PDA

View Full Version : Elo and PVP



Prodygi
08-23-2013, 09:03 PM
I'm curious to know if there be an Elo rating systems for Hex with regards to the PvP side of the game.

I certainly think it's a great idea to implement it into the game.

1) With the direction Hex is moving towards (E-sport, Twitch.tv), an elo system would help identify players of calibre. With a POWER LEVEL attached to a player, it would also give audiences some information of the player should he or she decides to stream.

2) It makes games more exciting, people try harder as there is a gain/lost with every victory or defeat.

3) An elo ranked system can provide players with PvP fun without any financial investment as there is no prize (boosters) attached to it. Just numbers.

4) This is a proven system. Notably with games and sports such as League of Legends and chess.



What do you think?

Apologies if similar post exists already.

Mr.Funsocks
08-23-2013, 09:07 PM
Many, many similar posts exist. And they've confirmed that there is some sort of rating system coming for ranked PvP. I *think* they mentioned Elo, but that could just be what's been suggested many times.

Prodygi
08-23-2013, 09:38 PM
Many, many similar posts exist. And they've confirmed that there is some sort of rating system coming for ranked PvP. I *think* they mentioned Elo, but that could just be what's been suggested many times.

Oops. In my defense, I did do a search before posting(couldn't find anything). If you could direct me to any similar post focusing on this subject or anything official by CZE it would be great. Or we can just post ideas here! =)

Kroan
08-24-2013, 12:57 AM
The only thing an ELO system will do is prevent people from wanting to play in tournaments once they hit a certain ELO. It didn't work in Magic, that's why they changed it. Not that their new system is the best solution, it's still better than ELO

The problem with ELO is that in chess a play with a high ELO will (or should) not lose to someone with a low ELO. This is because in Chess there is no luck involved. In Hex however there is a fair amount of luck and a reason why the game is appealing to skilled and nonskilled person. Once two players sit down, any of the two can win the match.

Shadowelf
08-24-2013, 01:00 AM
This is still a work in progress but they are considering ELO and a bunch of other different systems, like rankings for limited/constructed where you will gaining points that will eventually qualify you for tournaments like the world qualifier tournament (https://itunes.apple.com/us/podcast/...ign-mpt=uo%3D4 episode 54 check 0:40:25). Also Cory on organized play https://www.youtube.com/watch?v=f9mUXyn2jU4

Unhurtable
08-24-2013, 01:27 AM
The only thing an ELO system will do is prevent people from wanting to play in tournaments once they hit a certain ELO. It didn't work in Magic, that's why they changed it. Not that their new system is the best solution, it's still better than ELO

The problem with ELO is that in chess a play with a high ELO will (or should) not lose to someone with a low ELO. This is because in Chess there is no luck involved. In Hex however there is a fair amount of luck and a reason why the game is appealing to skilled and nonskilled person. Once two players sit down, any of the two can win the match.

Why would you not want to play in tournaments when you have high ELO?

Kroan
08-24-2013, 01:29 AM
Why would you not want to play in tournaments when you have high ELO?
Losing will lower your ELO (a lot if it's against players with low ELO) which might cause you to lose benefits that are bound to high ELO (invites to certain events for example). So once you hit a certain ELO, you just don't want to play anymore in that particular season. Which is very bad for the game for obvious reasons.

keldrin
08-24-2013, 01:38 AM
Losing will lower your ELO (a lot if it's against players with low ELO) which might cause you to lose benefits that are bound to high ELO (invites to certain events for example). So once you hit a certain ELO, you just don't want to play anymore in that particular season. Which is very bad for the game for obvious reasons.

This is why I favor, the concept of it being easier to gain points (by playing higher rated players) than to lose them. Essentially to help prevent people from being afraid to play. Another reason, being that if you have one deck that you perform really well with, limiting how quickly you can lose points, once your rating is established at that level (so many games playing similarly rated players with a decent win loss record kind of proving you really do play at that level), will make people feel a bit freerer about trying out new decks that they think have a good chance for decent wins.
Also, I think there should be a degrade of level, if you don't play for awhile. This would mean you have to play some, to maintain your rank level. not to mention, if you're really not playing, your game play will likely suffer.

Kroan
08-24-2013, 01:42 AM
Or, just an idea, don't use ELO system? You don't have to make ELO system work just to use ELO system... you can just aswell make a system where you acquire points during a season, and those points will qualify you for stuff. And the points are reset at the beginning of each season. Winning nets you points, losing does nothing.

There is no need to re-invent the wheel, magic went through the same systems, and ELO was in the end unfit and ditched in favor of the current system.

Edit: Here is a video by Evan Erwin that might explain it better than me during the announcement of PWP: http://www.youtube.com/watch?v=3Bqs6LPQCWI (Note that the system was refined the last few years, so mainly the first part is intresting)

Unhurtable
08-24-2013, 02:02 AM
Losing will lower your ELO (a lot if it's against players with low ELO) which might cause you to lose benefits that are bound to high ELO (invites to certain events for example). So once you hit a certain ELO, you just don't want to play anymore in that particular season. Which is very bad for the game for obvious reasons.

But the ELO wouldn't be lowered just because you lose in a tournament. As the OP mentions League of Legends as an example, the way it works there is that only matchmaking takes ELO into account. In other words, if you are invited to a "high ELO tournament" and lose a couple of games, your matchmaking ELO would still remain the same. In LoLs system there is also rating decay, if you don't play in a while your ELO is automatically lowered overtime. It is intensified for high-tier players, which I think have to play every couple of days to not lose their high rating.

Kroan
08-24-2013, 02:13 AM
But the ELO wouldn't be lowered just because you lose in a tournament. As the OP mentions League of Legends as an example, the way it works there is that only matchmaking takes ELO into account. In other words, if you are invited to a "high ELO tournament" and lose a couple of games, your matchmaking ELO would still remain the same. In LoLs system there is also rating decay, if you don't play in a while your ELO is automatically lowered overtime. It is intensified for high-tier players, which I think have to play every couple of days to not lose their high rating.
The ELO system was dropped by League of Legends in favor for the "League System" which works as you describe (from what I understand at least)

Btw; If you don't lose rating when losing, it's not an ELO-system.

Unhurtable
08-24-2013, 02:28 AM
The ELO system was dropped by League of Legends in favor for the "League System" which works as you describe (from what I understand at least)

Btw; If you don't lose rating when losing, it's not an ELO-system.

The ELO system works in the background, with the League System working as a system to give players goals. You can test this out by having a friend try out ranked for the first time with you having played roughly 50 games of Silver/Low Gold. If you win lots of games in a row, your friend might be beneath you in the League System but will show up above you in the ranked game lobby, indicating that he has higher rating than you (as the lobby positions are based on highest rating closest to first pick, with the exception that if a duo-team is present, the highest rated within the duo-team(s) will be first picker, which is the cause for the "first picker + last picker is duo" perception).

You lose rating when you lose in matchmaking, but as a tournament is not matchmaking, the ELO rating is not affected by wins/losses in tournaments.

Prodygi
08-24-2013, 03:29 AM
Elo and tournaments are two different things imo.

Elo is more for a gauge of how good you are as a player or, for a lack of better word, e-peen.
Tournament is a stage to prove your chops and prizes.

I actually believe that having an elo system encourages people to play the game. Everyone wants a higher number beside their name to prove how good they are right? Just look at the amount of "how to climb elo" or "get out of elo hell" articles out there for various games. Elo, while being only a number, motivates people to play.

LoL is a game with luck involve. Are you the blue or purple side? Did you pick first so the opponent can counter-pick you?
Hex has the luck element as well. So is that to say that whoever wins the tournament, won by luck? I don't think so.
As such, is the elo system IS accurate to a degree. Proven by the fact that the top players are always at the top. They have many different accounts and all of their accounts are right up there. It is an accurate measure of where one stands.


The ELO system works in the background, with the League System working as a system to give players goals. You can test this out by having a friend try out ranked for the first time with you having played roughly 50 games of Silver/Low Gold. If you win lots of games in a row, your friend might be beneath you in the League System but will show up above you in the ranked game lobby, indicating that he has higher rating than you (as the lobby positions are based on highest rating closest to first pick, with the exception that if a duo-team is present, the highest rated within the duo-team(s) will be first picker, which is the cause for the "first picker + last picker is duo" perception).

You lose rating when you lose in matchmaking, but as a tournament is not matchmaking, the ELO rating is not affected by wins/losses in tournaments.

Yup. The league system still holds the fundamental rule of the elo system.
You get matched with people of your elo.
You win, you gain points. You lose, you lose points.
Just that instead of numbers, they post you into brackets. But the fundamentals are pretty much the same tbh.

Kroan
08-24-2013, 04:57 AM
I've seen plenty of posts but none actually say anything about the upside of an ELO-system. Remember, there is no matchmaking announced and it's unlikely that there will be matchmaking to ensure all events fire as fast as possible.

Also a number to show how good you are is not an argument for ELO, since that also would be accomplished by just tracking wins, losses and percentage.


LoL is a game with luck involve. Are you the blue or purple side? Did you pick first so the opponent can counter-pick you?
Hex has the luck element as well. So is that to say that whoever wins the tournament, won by luck? I don't think so. That all may be very well true, however if you are a new player and you play against the national champion there is no way you can win in LoL but in Hex you actually have a shot at winning. Variance is a much bigger part of Hex than it is on LoL. Making an ELO system horrible for Hex.

Again; there is a reason why Magic ditched the ELO-system. Proplayers have been praying to get rid of the system so they actually can play Magic whenever they want. MTGO itself actually never used the ELO system (it has it built in, but just completly ignores it) and works with qualifier points.

Prodygi
08-24-2013, 05:25 AM
I've seen plenty of posts but none actually say anything about the upside of an ELO-system. Remember, there is no matchmaking announced and it's unlikely that there will be matchmaking to ensure all events fire as fast as possible.

Also a number to show how good you are is not an argument for ELO, since that also would be accomplished by just tracking wins, losses and percentage.


I believed I have mentioned a few.
When games or sports are designed to have a competitive nature, there will need to be a gauge for people to know where they stand. And an elo system accomplishes that and therefore a needed feature imo.
Tracking wins, losses and percentage wouldn't be very accurate. People would just prefer to play with people who's lousier than them. It gives NO reason for people to play with opponent better than them. Or one would just make 2 account and pump one up.
Games are more exciting/tryhard with the elo system given that there are stakes involve.
Elo promotes playing the game for a high number of the player base as it is an achievement/target they want to work towards.
Elo has NOTHING to do with events/tournaments. You get seeded to events or tournaments by winning events or tournament.
In the magic video you've linked, the top comment basically mirror my thoughts.

Your concern with people stop playing when they've reached a very high elo is valid. Lets agree that this problem is mainly for the top 10 percentile of the player base.
First solution is Elo-decay. You stop playing, you're going down.
Then in LoL, there is the Challenger tier. Which promotes playing from the cream of the crop to truely be the best.
And lastly, a Elo-Matchmaking queue is not the only PVP a player can play. If one is really in the top 10 percentile, they've still gotta stop their elo from decaying/reach challenger tier. And after that, they'll probably be better off going infinite in constructed given that they are so damn good.
Again, I stress that this system is an ADDITION to the other available form of PvP. Needless to say, elo would not be affecting other forms of PvP/matchmaking.

Kroan
08-24-2013, 05:50 AM
Tracking wins, losses and percentage wouldn't be very accurate. People would just prefer to play with people who's lousier than them. It gives NO reason for people to play with opponent better than them. Or one would just make 2 account and pump one up. I'm pretty sure that that won't work, since you'll have to enter a tournament and magically always be matched against "yourself". There is no "preference" you can use to play against "lousy" people. You enter a 8-man queue or a tournament and are matched randomly against people.

Games are more exciting/tryhard with the elo system given that there are stakes involve. Isn't the prize of the tournament in itself already something that's at stake?

Elo has NOTHING to do with events/tournaments. You get seeded to events or tournaments by winning events or tournament. If it has nothing to do with events/tournaments, then what is the point of having it at all? If you mean to track your "casual" games, than an ELO system makes even less sense.

Maybe I'm misunderstanding, but where are you actually going with this at all? Some kind of ladder-system where you join for random games? Isn't that already done with the "Marathon Tournaments"?

Prodygi
08-24-2013, 06:46 AM
This has NOTHING to do with events or tournaments. There is no entry fee, there is no prize. Just an increase or decrease of elo.

Does the rank system of LoL has anything to do with tournament? No. But then what does the elo system accomplishes?

"Games are more exciting/tryhard with the elo system given that there are stakes involve.
Elo promotes playing the game for a high number of the player base as it is an achievement/target they want to work towards.
With the direction Hex is moving towards (E-sport, Twitch.tv), an elo system would help identify players of calibre.
An elo ranked system can provide players with PvP fun without any financial investment as there is no prize (boosters) attached to it."

To add on elo is often a consideration for forming/recruitment into teams. Many top teams in LoL would not consider players below 2000 elo. But does elo achieve anything in the game of LoL, tournaments or events? Again, No.

Kroan
08-24-2013, 07:14 AM
So if it doesn't do anything at all. What's the point? LoL is not a valid example or reference case, it's totally different genre. MTGO has an ELO rating, which doesn't do anything and nobody bothers with it. It's completely useless and, afaik, removed in the next version.

Also, if you don't track ELO while playing sanction tournaments.... when do you actually track ELO then? I don't get it :confused:

Prodygi
08-24-2013, 07:51 AM
So if it doesn't do anything at all. What's the point?
Also, if you don't track ELO while playing sanction tournaments.... when do you actually track ELO then? I don't get it :confused:


what does the elo system accomplishes?

"Games are more exciting/tryhard with the elo system given that there are stakes involve.
Elo promotes playing the game for a high number of the player base as it is an achievement/target they want to work towards.
With the direction Hex is moving towards (E-sport, Twitch.tv), an elo system would help identify players of calibre.
An elo ranked system can provide players with PvP fun without any financial investment as there is no prize (boosters) attached to it."

To add on elo is often a consideration for forming/recruitment into teams.

I believe most of your arguements revolve around points that i've already answered (at least I think I did).

ramseytheory
08-24-2013, 08:41 AM
I'm really not sure why people want Elo here. The entire idea behind it is to use statistics to accurately estimate a player's skill based on their past performance. The model is that a player's performance is normally distributed (strictly speaking logistically distributed) around some mean - their "skill". Then player X beats player Y if and only if they perform better in the game. So if player X is believed to have far higher skill than player Y, but loses, this provides strong statistical evidence that player X has lower skill and/or player Y has higher skill than previously thought. You can use this to obtain estimates for people's skills - the Elo rankings.

But in Hex that model fails completely, since performance isn't the only factor. Even a world championship-level player will lose if they're mana-screwed badly enough, or if they're heavily restricted in their card choice (as new players are), so the statistical evidence obtained from X beating Y is dramatically reduced. I'm not a statistician so I'm not sure how you'd fix things, but I don't think tweaking parameters would be enough.

The upshot is that Elo is basically a broken model for Hex (although it works well for deterministic games like chess), so we should be looking for another model that does the same job - probably one that doesn't put too much weight on any single game. I agree Hex needs some Elo-like system, or newbies will be matched against people with tournament-level decks, but I don't understand why people are set on Elo specifically.

(As an aside, I think there should be at least an option for unranked play - otherwise people are going to find it very hard to test new decks against human players without endangering their ranking. And that's going to be an issue whatever model is used.)

Niedar
08-24-2013, 08:51 AM
I don't see how its a broken model. Someones "skill" should still be distributed around a mean, the only thing that is different is that there is a higher variance due to the randomness. I don't actually know the algorithm for Elo but I find it hard to believe that it can't take into account different variances.

Kroan
08-24-2013, 09:02 AM
I believe most of your arguements revolve around points that i've already answered (at least I think I did). You actually don't. Your whole point evolves around a system that tracks an elo rating on games that are played in some kind of matchmaking system, where Hex actually centers around limited and constructed tournaments. There has not been any word about any kind of matchmaking system to play constructed games, except some kind of lobby where you most likely can challenge people to play a casual game of Hex or test your deck for tournaments.

So not only is ELO not a good method to determine skill in tcg, it's also completely unneeded and irrelevant the way you describe it to be implemented.

If we actually want to track skill / points / whatever in tournaments (i.e. to qualify for tournaments) my original point still stands that you rather want to track wins for a certain time period (season)


I don't actually know the algorithm for Elo but I find it hard to believe that it can't take into account different variances. In ELO you have a rating. If you win against someone you gain points based on the rating difference. If you lose, you will lose points based on the rating difference. The problem is that if someone plays in a draft and just has bad luck and plays his first round against someone with a low rating he will lose a ton of points, even though it has nothing to do with his skill.

If you look at proplayers in Magic, a 60% win ratio in limited is actually considered outstanding.

Niedar
08-24-2013, 09:19 AM
So first of all that is not an algorithm but more of a general description and second of all how many points is awarded or deducted is obviously tunable which is why I said that for sure the actual mathematical algorithm for ELO has a factor in it that takes the variance into account. Losing or winning because of some inherent random conditions of the game is not really a problem except for the fact that it takes more total games played to accurately get any type of score.

Prodygi
08-24-2013, 10:00 AM
So are you saying that there should be no PvP apart from tournaments and events?
If not, can you offer a PvP system that doesn't involve money, doesn't require HOURS of commitment and has a stake?


So not only is ELO not a good method to determine skill in tcg, it's also completely unneeded and irrelevant the way you describe it to be implemented.
Still pretty sure my points are valid. But to each his own I guess.


If you look at proplayers in Magic, a 60% win ratio in limited is actually considered outstanding.
You'll be surprise how high a 60% win rate can bring you.


I agree Hex needs some Elo-like system, or newbies will be matched against people with tournament-level decks, but I don't understand why people are set on Elo specifically.

(As an aside, I think there should be at least an option for unranked play - otherwise people are going to find it very hard to test new decks against human players without endangering their ranking. And that's going to be an issue whatever model is used.)
Agreed. I think elo does that though. Suggestions?

The_Wine_Gnat
08-24-2013, 10:03 AM
I'm a HUGE fan of the League System (which uses ELO) for LOL. Went from Bronze 5 (the lowest) to Gold 3. Sure I've swung up and down alot, but it's fun knowing that I am improving. Better yet, I'm always matched against other players my skill level which keeps the level of fun very high. There really aren't that many stomps.

I would LOVE to see a league system (which uses ELO) for Hex. I want to play against other players of my caliber and have many close fun games. I do NOT want to get stomped by someone who has been playing CCGs their entire life. Those players (which I would stream), should play against players of their caliber so everyone can enjoy watching their bouts.

Kroan
08-24-2013, 10:12 AM
You'll be surprise how high a 60% win rate can bring you. You'll be surprised how a win rate of 60% will destroy any ELO rating you might have.

So first of all that is not an algorithm but more of a general description and second of all how many points is awarded or deducted is obviously tunable which is why I said that for sure the actual mathematical algorithm for ELO has a factor in it that takes the variance into account. ELO does NOT have a factor for variance. ELO is a system made for chess by a chessplayer. In chess there is no variance.

Losing or winning because of some inherent random conditions of the game is not really a problem except for the fact that it takes more total games played to accurately get any type of score This is not true. If this would be true, we would have Poker players ranked with ELO.

So are you saying that there should be no PvP apart from tournaments and events? What other PvP are you thinking of?

Anyway. I could go on and on about how ELO is not fit for Hex, but I think I made all points I can make and obviously I can't convince, even though Magic already used ELO for 16 years and ditched it because it was inherently bad for everyone involved.

zadies
08-24-2013, 10:35 AM
I wasn't aware that there was announced to be a non-paid match you against a bunch of ppl practice match que for this game and it was all draft tournaments of some sort.
Ummm they have been quite clear that PvP tournaments were how they were making all their money creating a system that encourages non-draft reward driven PvP play is actually counter productive to cze. Casual challenges in a lobby for practice does not detract from the draft tournaments a non-paid elo style que though would.

Unhurtable
08-24-2013, 11:19 AM
I'm really not sure why people want Elo here. The entire idea behind it is to use statistics to accurately estimate a player's skill based on their past performance. The model is that a player's performance is normally distributed (strictly speaking logistically distributed) around some mean - their "skill". Then player X beats player Y if and only if they perform better in the game. So if player X is believed to have far higher skill than player Y, but loses, this provides strong statistical evidence that player X has lower skill and/or player Y has higher skill than previously thought. You can use this to obtain estimates for people's skills - the Elo rankings.

But in Hex that model fails completely, since performance isn't the only factor. Even a world championship-level player will lose if they're mana-screwed badly enough, or if they're heavily restricted in their card choice (as new players are), so the statistical evidence obtained from X beating Y is dramatically reduced. I'm not a statistician so I'm not sure how you'd fix things, but I don't think tweaking parameters would be enough.

The upshot is that Elo is basically a broken model for Hex (although it works well for deterministic games like chess), so we should be looking for another model that does the same job - probably one that doesn't put too much weight on any single game. I agree Hex needs some Elo-like system, or newbies will be matched against people with tournament-level decks, but I don't understand why people are set on Elo specifically.

(As an aside, I think there should be at least an option for unranked play - otherwise people are going to find it very hard to test new decks against human players without endangering their ranking. And that's going to be an issue whatever model is used.)

You don't have to tweak the formula because overtime those with higher skill should still win most of the time, this is possible even in deterministic games. A better player can lose some games to worse players, but in most games the better player should win.


You'll be surprised how a win rate of 60% will destroy any ELO rating you might have.
ELO does NOT have a factor for variance. ELO is a system made for chess by a chessplayer. In chess there is no variance.
This is not true. If this would be true, we would have Poker players ranked with ELO.
What other PvP are you thinking of?

Anyway. I could go on and on about how ELO is not fit for Hex, but I think I made all points I can make and obviously I can't convince, even though Magic already used ELO for 16 years and ditched it because it was inherently bad for everyone involved.

How will having a high win rate destroy my ELO rating?
As mentioned above, ELO doesn't need a factor for variance.

Poker is not a centralized service, and since its most often played for money, ratings are unnecessary. We are talking about a system that is essentially zero reward.


I wasn't aware that there was announced to be a non-paid match you against a bunch of ppl practice match que for this game and it was all draft tournaments of some sort.
Ummm they have been quite clear that PvP tournaments were how they were making all their money creating a system that encourages non-draft reward driven PvP play is actually counter productive to cze. Casual challenges in a lobby for practice does not detract from the draft tournaments a non-paid elo style que though would.

Who said matchmaking would yield rewards? Because I recall:


3) An elo ranked system can provide players with PvP fun without any financial investment as there is no prize (boosters) attached to it. Just numbers

being posted in the OP.

Drafting will still remain the best option for players with boosters since you have the option to earn more boosters. Boosters would still remain the only way for PvP cards to be generated, which in other words would not be counter productive to CZE. It could even be argued that providing a matchmaking service would increase potential customers and keep current customers playing (as they no longer have to seek out matches themselves if they want a good challenge).

zadies
08-24-2013, 11:36 AM
Didn't say it would give rewards stating an e-peen rating system would decrease the number of ppl playing in tournaments thus cutting into the profit margin because you introduce an incentivized rating based f2p system into what is supposed to be the p2p portion of the game... Great for the consumer terible for cze.

Unhurtable
08-24-2013, 11:47 AM
Didn't say it would give rewards stating an e-peen rating system would decrease the number of ppl playing in tournaments thus cutting into the profit margin because you introduce an incentivized rating based f2p system into what is supposed to be the p2p portion of the game... Great for the consumer terible for cze.

Why would people stop playing in tournaments just because there is a matchmaking system in place?
1. Tournaments would still be more prestigeous than matchmaking.
2. Tournaments would still be more rewarding than matchmaking (as there are rewards in tournaments).
3. There will already be PvP outside of tournaments and drafts. All that is being asked for is to make it simpler to play versus people of the same skill-level. You also have nothing to back up the claim that the number of people playing in tournaments would decrease if a simple matchmaking system was put in place.

zadies
08-24-2013, 03:43 PM
People either play for prestige or rewards if you track prestige separately from rewards you split the population. Also if given recognition for something free vs something you pay for some subset of PvP players will elect yo do the free thing. Also since you can't do both at once making it easier to do for free where the ranked completion is going to be more appealing to a certain subset of players.
Or are you saying if this was implemented then you wouldn't choose to play this over the draft tournaments?
Or are you saying you wouldn't participate in PvP at all if this wasn't done?
Because if the answer to the the first question is yes and you would be participating in PvP without a free elo ranking system then why would they want to cut into their profits, but if you can honestly answer no you would not draft less and feel that every single PvP player in the world would not draft less if there was a free show how badass you are option then it might be worthy of continuing this discussion. If playing for free nets you nothing not even bragging rights that is one thing but taking the time to design a que system that nets no profit and in fact will likely decrease the number of people drafting and splitting the PvP base makes no sense especially for a system that has proven not to work in other tcgs.
Now if you wanted to discuss paying a premium to be able to play in some sort of style ladder that lasted the length of a set that might be a possibility but the skill to luck ratio in this game is much more in favor of luck the LoL which is one of the many reasons elo is not a good idea and why decisions are based 2out of 3 in tcgs and not so in games where skill is the greater determinate.

ramseytheory
08-24-2013, 03:44 PM
I don't see how its a broken model. Someones "skill" should still be distributed around a mean, the only thing that is different is that there is a higher variance due to the randomness. I don't actually know the algorithm for Elo but I find it hard to believe that it can't take into account different variances.

The problem is that Elo assumes performance (which here depends on both skill and luck) is normally distributed, and if you take that away it's useless. The reason for the assumption is that if a player's performances in different games are independent (mostly valid) and identically distributed (kind of valid-ish as long as players' skills don't increase too quickly), then after a long enough time their average performance is guaranteed to resemble a normal distribution by the central limit theorem. The issue is that IIRC "a long enough time" is far longer when your variables have high variance, and to account for maybe a 10% chance of an excellent player losing to an awful player you'd need to use very high variance indeed.

So basically, if you just took Elo and increased the variance, you might get an accurate estimate of people's rankings over a hundred million games, but you probably wouldn't get one over 50 games. I'm sure it's a fixable problem - I certainly don't think a working matchmaking system is impossible - but just taking vanilla Elo and tweaking the variance of the normal variables seems likely to fail.

(I'm ignoring logistic variables here since I know very little about them...)

zadies
08-24-2013, 04:27 PM
Other then some people wanting to brag about a completely meaningless number due the number of games it would take to provide an accurate number to avoid the variance that luck injects into the equation. Yes a 60% win record is great until you factor in the fact that means the world top competors still lose 40% of the time to ppl with sub 1000 scores which elo predicts should win 2% of the time if the scores are actually accurate.

Prodygi
08-24-2013, 08:54 PM
Having an ADDITIONAL PvP system that doesn't require an entry commitment and has some form of prestige attached to it will increase sales for cze rather than the other way round.

1) It attracts more people to dabble into PvP. If there is an entry fee for all forms of PvP, can you imagine how much of the player base will not touch PvP at all? With games, player base = profits. Which is why CZE introduces a COMPLETELY F2P PvE side of the game. Because they understand that simple fact. My idea accomplishes that. It attracts people to try, to take a first step in PvP.

2) With a stake involve, people play to win. In Hex, that means upgrading your deck. That means buying/trading of cards. That will generate activities and involvement for the player towards the game, and profits for CZE. Some players will simply buy the cards from the AH, but some will go into drafting to expend their deck quality. See, you basically just introduce drafting to a "free" player. Also, with different forms of PvP to play, it keeps the game fresh and not repetitive and thus increasing it's playability.

3) There NEEDS to be some form of free pvp system with some form of stake. Seriously, paying $5 everytime a player wants to PvP is NOT the way to go. There will be players who wants to PvP but do not need anymore cards. At this point, which PvP queue can they go? If there are no stakes involve, this scenario will happen:
Player A: Crap draw. Mulligan to 6. Crap again, mulligan to 5. Damn!
Player A: Ok, GG you win. /quit
Player B: Wtfbbq?

Elo rating system is a system that fits this ideal. But lets not tunnel-vision into that. Because end of the day, it can be MMR, the hearthstone ranking system, etc, I don't really care. What I'm suggesting is that there needs to be a system that:
1) Matchmake players of equal skills to play one another
2) Free but has some form of reward/punishment

Edswor
08-25-2013, 12:42 AM
Hex must have a system to rank players "internally" (as we don't need to see the number) so there can be a good matchmaking system (aside from tournaments, etc). As Prodygi point out, it doesn't matter if its ELO, MMR or other system as long as it provides a ranking system for good matchmaking.

Unhurtable
08-25-2013, 02:04 AM
People either play for prestige or rewards if you track prestige separately from rewards you split the population. Also if given recognition for something free vs something you pay for some subset of PvP players will elect yo do the free thing. Also since you can't do both at once making it easier to do for free where the ranked completion is going to be more appealing to a certain subset of players.
Or are you saying if this was implemented then you wouldn't choose to play this over the draft tournaments?
Or are you saying you wouldn't participate in PvP at all if this wasn't done?
Because if the answer to the the first question is yes and you would be participating in PvP without a free elo ranking system then why would they want to cut into their profits, but if you can honestly answer no you would not draft less and feel that every single PvP player in the world would not draft less if there was a free show how badass you are option then it might be worthy of continuing this discussion. If playing for free nets you nothing not even bragging rights that is one thing but taking the time to design a que system that nets no profit and in fact will likely decrease the number of people drafting and splitting the PvP base makes no sense especially for a system that has proven not to work in other tcgs.
Now if you wanted to discuss paying a premium to be able to play in some sort of style ladder that lasted the length of a set that might be a possibility but the skill to luck ratio in this game is much more in favor of luck the LoL which is one of the many reasons elo is not a good idea and why decisions are based 2out of 3 in tcgs and not so in games where skill is the greater determinate.

You are assuming that the population will be splitted by this implementation, rather than the (in my opinion) more reasonable expectation that the population will participate in both activities.
Once again, people WILL do non-draft, non-tournament PvP even without a matchmaking system in place, which will still be able to cause recognition (as you might duel a big-shot player and win). Do you want to remove private matches as well?
The people who would be more drawn to matchmaking (and less drawn to draft, for whatever reason) would still do PvP outside of draft / tournaments.
I would still play draft / tournaments if a matchmaking system was in place. Tournaments are more prestigeous (and grant rewards) and drafts would still remain the best option to spend your boosters on. Why can my side of the argument only be logical / worthy of discussion if every single PvP player in the world would do as I? If anything, more exposure to cards / decktypes will boost sales in boosters, which leads to more drafts being done. Also, I can guarantee you one thing. People won't spend money they don't have on drafts just because there is no other option for PvP.
I would still participate in PvP even if this wasn't implemented. But I don't think I would do more drafts / tournaments just because there is no matchmaking.
will likely decrease the number of people drafting and splitting the PvP base which I will just counter with "will likely increase the number of people drafting and uniting the PvP base". Neither of us have anything to back up our claims.
Actually, the skill to luck ratio is worse in solo-queue LoL since 80% of your rating is essentially based on something you almost cannot control (your teammates). I'd love to see someone argue that 80% of a MTG/Hex game is based on luck.

I must have said this like 10 times on this forum. Very deterministic games also use Bo3 and Bo5 matches. Its not something MTG or Hex does / will do because "the game contains random elements", but because worse teams / players CAN win versus better teams / players, but shouldn't in most games. Therefore Bo3 and Bo5 formats are great because you are truly the better team if you win most of the games.


The problem is that Elo assumes performance (which here depends on both skill and luck) is normally distributed, and if you take that away it's useless. The reason for the assumption is that if a player's performances in different games are independent (mostly valid) and identically distributed (kind of valid-ish as long as players' skills don't increase too quickly), then after a long enough time their average performance is guaranteed to resemble a normal distribution by the central limit theorem. The issue is that IIRC "a long enough time" is far longer when your variables have high variance, and to account for maybe a 10% chance of an excellent player losing to an awful player you'd need to use very high variance indeed.

So basically, if you just took Elo and increased the variance, you might get an accurate estimate of people's rankings over a hundred million games, but you probably wouldn't get one over 50 games. I'm sure it's a fixable problem - I certainly don't think a working matchmaking system is impossible - but just taking vanilla Elo and tweaking the variance of the normal variables seems likely to fail.

(I'm ignoring logistic variables here since I know very little about them...)

A hundred million games require HUGE amounts of variance. Think dice rolls. Hex is not a game of dice rolls.


Other then some people wanting to brag about a completely meaningless number due the number of games it would take to provide an accurate number to avoid the variance that luck injects into the equation. Yes a 60% win record is great until you factor in the fact that means the world top competors still lose 40% of the time to ppl with sub 1000 scores which elo predicts should win 2% of the time if the scores are actually accurate.

Source? I'd love to see statistics of the players above 1800 losing 40% of their games to people sub 1000. Because as far as I know, the 60% win record is a mix between all the ratings. You maybe win 20% of your games versus 2k players, 50% versus 1.8k players, 80% versus 1.2k players and 95% versus 1k players.

ramseytheory
08-25-2013, 05:02 AM
What worries me if a ranking system is used for all PVP is not that people won't play tournaments or draft at all, but that some people won't play tournaments or drafts unless they're on top form. I can very easily see competitive players being more reluctant to decktest in constructed tournaments, to draft an experimental deck, to raredraft, or to play any PVP while tired or ill, if doing so was likely to reduce their rankings.



A hundred million games require HUGE amounts of variance. Think dice rolls. Hex is not a game of dice rolls.


An Elo 2800 player has maybe a 10% chance of losing against an Elo 1000 player (given equally powerful decks). If we're trying to model the game the Elo way, this means your normal variables would need standard deviation maybe 1200 or so, so variance maybe 1.4 million. That's quite large.

Prodygi
08-25-2013, 05:18 AM
If the said people wouldn't play a free PvP, why would they throw money to PvP at all?
Also, the rating system is not used for all pvp. It is an independent queue that only affects itself. Meaning the ratings would hold no effect should one queue for constructed(the one with the entry fee and prizes) or drafting.

Why would the 2800elo player be using the same deck as a 1000elo player?
And if they're both using the same deck and the 1000elo player is at where he is, it says a lot about the 1000elo player don't you think?

Unhurtable
08-25-2013, 05:53 AM
What worries me if a ranking system is used for all PVP is not that people won't play tournaments or draft at all, but that some people won't play tournaments or drafts unless they're on top form. I can very easily see competitive players being more reluctant to decktest in constructed tournaments, to draft an experimental deck, to raredraft, or to play any PVP while tired or ill, if doing so was likely to reduce their rankings.

Which will entirely depend on how they will do drafts / tournaments. Also, the ranking would only be affected by matchmade games, not drafts / tournaments.


An Elo 2800 player has maybe a 10% chance of losing against an Elo 1000 player (given equally powerful decks). If we're trying to model the game the Elo way, this means your normal variables would need standard deviation maybe 1200 or so, so variance maybe 1.4 million. That's quite large.

Wat?

zadies
08-25-2013, 09:07 AM
The issue is that elo can't predict the out come of a 1000 vs 2000 with any accuracy at all even in a mirror match. It comes down to card draw and in a game that uses elo should be predictive on a 1 time match not any kind of best of scenario used in tcgs. Tcgs use a best of system to try to cut down on variance and make it predictive of player skill but it does not eliminate it otherwise the top players in the world wouldn't feel 60% win percentage was good in draft.
Tcgs try to get around the card drawing variance by making them best of games, but for elo to be truely predictive in a tcg first off the number of games it would take prior to making any predictions would be huge, and we would need to move at minimum to a best of seven format like basketball, which in fact why basketball's finals were moved to best of seven so that there would be less upset wins. Chess does not use a best off rule because elo is highly predictive of the outcome, is the system becomes less predictive you start running best of series to try to eliminate that varriance.
Also splitting the PvP players between a ranked que and a prize que. With two ques you are splitting the population. Unless all players in the ranked que would not have participated in the prize que you are also lowering profits.
Splitting the population before you know its size into an additional que is also unwise because you will increase due times unless the population increase from adding the que covers all individuals that are using the que which you have already said it would not because you predicted that individuals would use both Que's(which infers unless there is enough population growth caused by the addition of the que then there are less people in the prize que.)
Also given the elo que you are proposing is a constructed que you can not be comparing it you are comparing its use against the draft tournament que not the constructed tournament que like you should be. I think it may lower the draft due marginally given it would be free but it would have a much larger impact on the constructed prize que where you are splitting the population between epeen and actual prizes. Given they already have a constructed que making another will decide that population.
Also due to the fact that an intelligent individual will use the ah to acquire most of the cards needed for their constructed deck rather then opening booster(though they could do this through drafting unless they are winning the draft tournaments in most cases it will likely be cheaper to ah the deck) after booster looking for a chase card I thought I would lay out why the use of the ah is not actually profit for cze which is why there should be an entry fee on the constructed tournaments.
Unless there is a large ah fee using plat as a medium of exchange is not profit for cze. Until it is used to purchase something directly from cze it is currently just a one way currency exchange just like if you visit a foreign country and change the currency from your national currency to the foreign nations the only stipulation is currently you can't change it back, but the simple conversion of the money does not directly profit anyone in the foreign country until an actual purchase is made so that the money is no longer in the circulation of consumers. Buying something off the ah keeps the currency in consumer hands less the ah fee, so only the ah fee is actual profit. The convince of the ah will likely cause more people to play hex which is a great thing, but don't confuse the idea that it is directly profiting them having a larger user base is why it is there.

Unhurtable
08-25-2013, 10:52 AM
The issue is that elo can't predict the out come of a 1000 vs 2000 with any accuracy at all even in a mirror match. It comes down to card draw and in a game that uses elo should be predictive on a 1 time match not any kind of best of scenario used in tcgs. Tcgs use a best of system to try to cut down on variance and make it predictive of player skill but it does not eliminate it otherwise the top players in the world wouldn't feel 60% win percentage was good in draft.


I must have said this like 10 times on this forum. Very deterministic games also use Bo3 and Bo5 matches. Its not something MTG or Hex does / will do because "the game contains random elements", but because worse teams / players CAN win versus better teams / players, but shouldn't in most games. Therefore Bo3 and Bo5 formats are great because you are truly the better team if you win most of the games.
Also, the ELO system is not designed to be able to predict single games. Its designed to predict odds of winning. People of higher rating CAN lose to people of lower rating, which is a part of the process of reaching "your true rating" which you will essentially never stay at, but move around, since your rating is essentially updated after each game.


Tcgs try to get around the card drawing variance by making them best of games, but for elo to be truely predictive in a tcg first off the number of games it would take prior to making any predictions would be huge, and we would need to move at minimum to a best of seven format like basketball, which in fact why basketball's finals were moved to best of seven so that there would be less upset wins. Chess does not use a best off rule because elo is highly predictive of the outcome, is the system becomes less predictive you start running best of series to try to eliminate that varriance.

Not really, a bo1 format would work well in matchmaking, as a bo3 might be a bit too long. Also, chess finals are not bo1. In the case of world championships, bo12 is the current format afaik.


Also splitting the PvP players between a ranked que and a prize que. With two ques you are splitting the population. Unless all players in the ranked que would not have participated in the prize que you are also lowering profits.

No, since the increase in PvP play (stemming from ease of access) would increase profits by incentivizing booster purchase (which in turn incentivizes draft).


Splitting the population before you know its size into an additional que is also unwise because you will increase due times unless the population increase from adding the que covers all individuals that are using the que which you have already said it would not because you predicted that individuals would use both Que's(which infers unless there is enough population growth caused by the addition of the que then there are less people in the prize que.)

Queue times stem from population and matchmaking. As drafts are not matchmade, queue times are directly correlated to the pool of players wanting to draft. As soon as there are 8 players in a draft pool, a new drafting session begins. For arguments sake, lets assume a normal draft queue (in a system without a free matchmaking) takes 10 minutes. We then introduce a free matchmaking system, and half of the current people who want to draft go to the free matchmaking instead. The queue would now be 20 minutes instead. I'd argue the problem began at the 10 minute draft queues, but what do I know, not to mention the actual impact could lead to anywhere between an additional minute to unable to find anything. Here are the questions we are trying to answer:
1. Will a free matchmaking option lead to more or less boosters being purchased per player? You say no I say yes.
2. Will a free matchmaking option lead to players playing Hex longer? I'd say yes.
3. Will a free matchmaking option lead to more players joining the community? I'd say yes.
Lastly, I'd combine these three answers and argue that CZE would economically benefit (more players + more boosters + players playing longer > longer draft queues).


Also given the elo que you are proposing is a constructed que you can not be comparing it you are comparing its use against the draft tournament que not the constructed tournament que like you should be. I think it may lower the draft due marginally given it would be free but it would have a much larger impact on the constructed prize que where you are splitting the population between epeen and actual prizes. Given they already have a constructed que making another will decide that population.

The people striving for epeen would still want to play tournaments. I can guarantee you that matchmaking will never be the highest prestige to be acquired in Hex, unless it gives PHAT rewards which nobody is arguing for. I can guarantee this mainly because there is no history of this ever happening in any other sport / game. I know, I know, argument from history.


Also due to the fact that an intelligent individual will use the ah to acquire most of the cards needed for their constructed deck rather then opening booster(though they could do this through drafting unless they are winning the draft tournaments in most cases it will likely be cheaper to ah the deck) after booster looking for a chase card I thought I would lay out why the use of the ah is not actually profit for cze which is why there should be an entry fee on the constructed tournaments.
Unless there is a large ah fee using plat as a medium of exchange is not profit for cze. Until it is used to purchase something directly from cze it is currently just a one way currency exchange just like if you visit a foreign country and change the currency from your national currency to the foreign nations the only stipulation is currently you can't change it back, but the simple conversion of the money does not directly profit anyone in the foreign country until an actual purchase is made so that the money is no longer in the circulation of consumers. Buying something off the ah keeps the currency in consumer hands less the ah fee, so only the ah fee is actual profit. The convince of the ah will likely cause more people to play hex which is a great thing, but don't confuse the idea that it is directly profiting them having a larger user base is why it is there.

I'm drawing a huge blank here. Are you saying a bigger population is not beneficial if there isn't an AH fee?

Niedar
08-25-2013, 11:00 AM
Either way, he is incredibly wrong. In order to buy something off the AH as a new player coming to the game, platinum has to be bought from Cryptozoic, that is an instant and direct profit for them. The only time what he said would be true is if I buy platinum from someone else on a secondary market. So I make an exchange with someone on paypal or something and then they trade me platinum in game.

Hollywood
08-25-2013, 11:29 AM
In tournaments, matchmaking will never use an ELO system but the normal Swiss pairings system.

An ELO ranking system is good for players who like to have the edification of their play skill that comes with a high rating. However, using it for tournament invites and/or rewards does pose the problem of player participation drop-out. A threshold system more akin to Magic's Planesealker Points or WoW's Honor Ranking system is better for this since it encourages player participation as opposed to discouraging it.

The real question is, why not just have both? The only negative that there is for implementing and tracking both systems is purely a matter of server space and man power. Both of which shouldn't be a problem if the Kickstarter support is any indication.

zadies
08-25-2013, 12:13 PM
Either way, he is incredibly wrong. In order to buy something off the AH as a new player coming to the game, platinum has to be bought from Cryptozoic, that is an instant and direct profit for them. The only time what he said would be true is if I buy platinum from someone else on a secondary market. So I make an exchange with someone on paypal or something and then they trade me platinum in game.
No I'm not wrong in the fact that cze still owes on the platinum saying that cze currently isn't allowing for exchange back when they have said that in a perfect world they would like to and also the fact that there will likely be some against EULA third party site that will likely come to be that buys the plat off you unless cze makes direct trading impossible. If the company still owes a service until said service is provided they really have made no profit.
The issue here is this really is in the end creates a system that attempts to minimize the already conceived of plat sinks. That may not be the original goal of the proposal but it is a conclusion from it. The pve is the hook to get ppl to PvP which is not nor have an endorsed f2p component.
I had already said the idea was fine if there was a participation fee. The portion of the game that is free is the pve it is likely going to be completely possible to make a pauper deck by just grinding gold, and a completely f2p component that is endorsed by cze is what is being asked for here.
Elo is not the right system to be using for match making though given the amount of games needed to make it even reasonably accurate LoL requires that you get get to lv 30 which is ~350 games and takes ~50 games before it lowers you elo change afterward. LoL has much less luck involved especially in arranged matches then Hex will...so it would likely take a dedicated player a year before elo became remotely useful in matchmaking.

Niedar
08-25-2013, 01:27 PM
They don't owe anything, platinum is not a currency it is a digital item that you don't even own and can be taken away from you at any time.

zadies
08-25-2013, 02:29 PM
You just said it was a currency and what do you do with currencies you spend them. The fact there is no direct means of changing it back to real money that is used to purchase things not provided by cze does not mean you don't exchange it fo r products and services does mean until the exchange is made they do owe you a product or services.

Niedar
08-25-2013, 05:35 PM
No, you just don't get it. Platinum is the product, its also a product that you don't actually own. You are not depositing money with CZE you are buying a digital item, it just so happens that the digital item can be used to buy other ingame digital items or traded with other players.

zadies
08-25-2013, 06:38 PM
I disagree given while you still have plat in your account you have 0 reason to spend more real money it just exists as a medium of exchange until it is turned in for something. So anything that elminiates/avoids a plat sink is really just existing to cut into cze's profit. Even if you want to say that the platinum is the product while you still have enough platinum to do whatever it is you are intending to do there is no reason for you to buy plat as a product so either way it a free endorsed way of doing pvp it cuts into thier profit regardless of wither you are considering the plat or the plat fee/boosters as where they get their profit.

Niedar
08-25-2013, 07:41 PM
Of course there is a reason to buy plat, so you can buy shit off the AH or buy packs or enter into tournaments. How do I come to have plat in my account in the first place? The majority of the time because I bought it from CZE, there will be a minority of cases where someone will be getting any large amount of play by PVE grinding and selling shit on the AH for plat. The only other way is to buy platinum from another player which again are the minority of cases.

What you are saying though is that if CZE wants to max their profit maybe they should just steal it all back from me, I mean then I would have to buy plat again.

What you can't seem to understand is having a free method of playing against other players does not necessarily mean less platinum sales, because maybe more people will be interested in playing a game where every time they want to play against another player they don't have to pay up. With that increased number of people playing and buying cards off the AH and buying booster packs it covers whatever was lost.

zadies
08-25-2013, 07:44 PM
That doesn't address the fact that not having plat sinks such as entry fees in pvp or buying boosters means that there is less need to buy plat which would mean that cze makes less profits implementing a f2p version of pvp.

Prodygi
08-25-2013, 07:53 PM
Here are the questions we are trying to answer:
1. Will a free matchmaking option lead to more or less boosters being purchased per player? You say no I say yes.
2. Will a free matchmaking option lead to players playing Hex longer? I'd say yes.
3. Will a free matchmaking option lead to more players joining the community? I'd say yes.
Lastly, I'd combine these three answers and argue that CZE would economically benefit

This.

I say again, having to pay money every time a player wants to PvP is not the way to go. It is not a plat sink, it's a turn off.
A good game should never put profit before consumer. A good product along with a good player base WILL result in profit.
Lets look at the MMOs out there that launched with a pay-to-play/pay-to-win(profit before consume) model.
Rift, Dungeons and Dragons Online, Tera, Lord of the Ring Online, Team Fortress 2... I could go on.
None of them are pay to play today. While their gameplay remains the same, ALL of them made more profits. Simply due to an increase in player base.
But what you are suggesting is, "Make PvP pay to play only so CZE can make more profits." I argue that that is wrong. Limiting PvP to pay to play WILL decrease PvP player base and as such, result in less profit for CZE.


That doesn't address the fact that not having plat sinks such as entry fees in pvp or buying boosters means that there is less need to buy plat which would mean that cze makes less profits implementing a f2p version of pvp.

1) No one said anything about removing existing queues with an entry requirement.
2) We have already argued that having an additional free PvP option will result in more players participating in PvP and thus increasing boosters' revenue.

zadies
08-25-2013, 08:02 PM
I'm sorry but Dugeons and dragons online and lord of the rings are quite pay to win... and if you don't think raid timer unlocks are just the tip of the pay to win iceberg there.
Also I'm not arguing make pvp profit only I'm arguing keep pvp as a plat sink which is how it was desigined.
Half the game is f2p they have been quite clear as to what where the profit was coming from and that was tournament fees. So saying that making pvp pay to play will decrease the player base anyone following the game since launch came into it knowing was pay to play thus your really arguing to increase the player base but wither this will increase profits is quite unknown given the f2p crowd already has pve to play. The f2p bait already exists there is no reason to add additional bait to the hook.

Niedar
08-25-2013, 08:13 PM
PvP already is pay to play, you can't pvp without cards, I shouldn't have to pay again to actually play with the cards I spent money on. The reason tournaments cost money to play is because there are rewards for winning but tournaments don't have to be the only way to play the game.

zadies
08-25-2013, 08:22 PM
Elo is not a good matchmaking system for it would take years for a casual player to have an accurate rating. Guild tournaments PvP team games and random challenges are alternate ways to PvP there is no reason to have them create a free matchmaking system that competes with the tournaments.

Prodygi
08-25-2013, 08:34 PM
"I'm arguing keep pvp as a plat sink which is how it was desigined."
And we rebutted that having a free PvP option will result in more players, more decks, more cards needed and therefore, more plat spent/profit.

"they have been quite clear as to what where the profit was coming from and that was tournament fees."
We are not saying to remove tournament fees. In fact, with a free PvP option, MORE people would be joining tournaments. (Due to a higher player base.)

"your really arguing to increase the player base but wither this will increase profits is quite unknown"
I've listed examples of games going free to play and profiting. Maybe you can name a few games that went pay to play and increasing it's revenue?

jimbeem
08-25-2013, 08:42 PM
Elo is not a good matchmaking system for it would take years for a casual player to have an accurate rating. Guild tournaments PvP team games and random challenges are alternate ways to PvP there is no reason to have them create a free matchmaking system that competes with the tournaments.

It doesn't take years to get an accurate ELO, it takes literally a few games. You either win and go up or lose and go down, either way until the reverse starts happening you know you are moving towards your correct rating. Also matchmaking does not, in any way, compete with a paid entry tournament with prizes for the victors. People shouldn't have to be in a tournament to pvp, they shouldn't have to pay every time they feel like playing a few games against players. matchmaking makes sense, anything else does not.

Chark
08-25-2013, 08:44 PM
The real question is, why not just have both? The only negative that there is for implementing and tracking both systems is purely a matter of server space and man power. Both of which shouldn't be a problem if the Kickstarter support is any indication.

The current school of thought in games is to no longer display a numeric value of your skill relative to other players. It encourages pretty toxic behavior. For example, magic online used to display Elo rating to players and that led to some pretty abusive interactions in games ("you're a 1600 scrub, I can't believe you're beating me, etc.").

Elo is a pretty bad indicator of individual performance in TCGs, because it was designed for a game with perfect information and no luck. In addition, given how TCGs normally ignore ratings in a tournament and pair people solely based on tournament performance (with the exception of R1). It doesn't make sense that a 2000-rated player could be dropped into a draft pod with 7 1000-rated players, lose 3 times and drop 30-40 points in rating. This person didn't have an aneurism--their true skill didn't decrease as much as the rating indicates. They just hit a streak of bad luck (probably the draft didn't go well for them).

Gliko-2 is a better rating system, although I still wouldn't expose it to players. If I were designing this, I would be looking at what starcraft 2 or league of legends are doing.

But to be honest, there isn't much demand for match made play. You are either:

1. Playing in a tournament for prizes, in which case, you don't want to be matched evenly. You always want a chance at lower rated opponents.
2. You are playing a heads-up match for a booster pack. See 1.
3. You are playing for fun. In that case you definitely don't want a rating system attached to those matches. It dis-incentivizes people trying new things and forces people to stick to the current metagame or risk tanking their rating.
4. You are testing for a major tournament. This is probably the only time where you want to be matched up against opponents of similar skill (since you assume that they are also testing with relevant metagame decks). However this function is just as easily fulfilled with a team (guilds). In fact it's strictly better to test with a team, to not give away anything interesting you develop in testing.

zadies
08-25-2013, 08:57 PM
It doesn't take years to get an accurate ELO, it takes literally a few games. You either win and go up or lose and go down, either way until the reverse starts happening you know you are moving towards your correct rating. Also matchmaking does not, in any way, compete with a paid entry tournament with prizes for the victors. People shouldn't have to be in a tournament to pvp, they shouldn't have to pay every time they feel like playing a few games against players. matchmaking makes sense, anything else does not.
It does with the amount of variance a tcg introduces.... It takes 350 games to even unlock ranked play in LoL and another 50 to have it lower the gain/loss per match... You would have to literally play thousands of games to try to a just for the variance caused by luck... Yes you can have an inaccurate number much faster but it wouldn't be an elk rating basically you are supposed to be 10x more likely to win against someone per 400 pts and given anyone can win on a given day elo just isn't a good system.
It really seems to me that people supporting introducing a ranked matchmaking que are doing so because it was proposed to be free not because a ranked que is a good idea in and of itself given no one has actually said they would pay for one.

dogmod
08-25-2013, 09:06 PM
The current school of thought in games is to no longer display a numeric value of your skill relative to other players. It encourages pretty toxic behavior. For example, magic online used to display Elo rating to players and that led to some pretty abusive interactions in games ("you're a 1600 scrub, I can't believe you're beating me, etc.").

Elo is a pretty bad indicator of individual performance in TCGs, because it was designed for a game with perfect information and no luck. In addition, given how TCGs normally ignore ratings in a tournament and pair people solely based on tournament performance (with the exception of R1). It doesn't make sense that a 2000-rated player could be dropped into a draft pod with 7 1000-rated players, lose 3 times and drop 30-40 points in rating. This person didn't have an aneurism--their true skill didn't decrease as much as the rating indicates. They just hit a streak of bad luck (probably the draft didn't go well for them).

Gliko-2 is a better rating system, although I still wouldn't expose it to players. If I were designing this, I would be looking at what starcraft 2 or league of legends are doing.

But to be honest, there isn't much demand for match made play. You are either:

1. Playing in a tournament for prizes, in which case, you don't want to be matched evenly. You always want a chance at lower rated opponents.
2. You are playing a heads-up match for a booster pack. See 1.
3. You are playing for fun. In that case you definitely don't want a rating system attached to those matches. It dis-incentivizes people trying new things and forces people to stick to the current metagame or risk tanking their rating.
4. You are testing for a major tournament. This is probably the only time where you want to be matched up against opponents of similar skill (since you assume that they are also testing with relevant metagame decks). However this function is just as easily fulfilled with a team (guilds). In fact it's strictly better to test with a team, to not give away anything interesting you develop in testing.

You make excellent points but you 1-4 only apply to pro or competitive players. Those points don't apply to a casual audience.

stiii
08-25-2013, 09:32 PM
The current school of thought in games is to no longer display a numeric value of your skill relative to other players. It encourages pretty toxic behavior. For example, magic online used to display Elo rating to players and that led to some pretty abusive interactions in games ("you're a 1600 scrub, I can't believe you're beating me, etc.").



Well to be fair you still get this, people will just assume the guy they are losing to is some scrub. Not that having a 1900 ranking stopped people complaining either.

stiii
08-25-2013, 09:33 PM
You make excellent points but you 1-4 only apply to pro or competitive players. Those points don't apply to a casual audience.

Are you saying casual players want a ranking system?

Prodygi
08-25-2013, 09:37 PM
Why not? It matches them to people of similar skill level.

zadies
08-25-2013, 10:03 PM
"I'm arguing keep pvp as a plat sink which is how it was desigined."
And we rebutted that having a free PvP option will result in more players, more decks, more cards needed and therefore, more plat spent/profit.

"they have been quite clear as to what where the profit was coming from and that was tournament fees."
We are not saying to remove tournament fees. In fact, with a free PvP option, MORE people would be joining tournaments. (Due to a higher player base.)

"your really arguing to increase the player base but wither this will increase profits is quite unknown"
I've listed examples of games going free to play and profiting. Maybe you can name a few games that went pay to play and increasing it's revenue?

No you have given examples of pay to win games profiting not f2p games and until trion releases profit numbers your assuming that rift is more profitable now then beforehand which was the only game mentioned that has a truly f2p model it was also the only game listed that wasn't nearly dead before going f2p. It is very easy to say f2p makes more money then shutting down. Also given trion just fired everyone adamant about trying to keep the game from being pay to win not sure how long that one won't be.
Saying more boosters will be opened when you can get a fair match out of a pauper deck from the ah given you will end up being matched with ppl with similar decks/skills means there really isn't an incentive for someone looking for free casual PvP to spend money which is the kind of ppl this would really target to get into PvP.

Prodygi
08-25-2013, 10:22 PM
You are assuming that people would happily stay where they are, instead of investing and trying to improve or up their elo/rank/bracket etc. One which I find highly unlikely.

What we are ultimately debating is my believes against yours.
I believe that having a bigger player base in PvP means more profit. You don't.

zadies
08-25-2013, 10:30 PM
I see the people who matchmaking with no prizes would appeal to would care about the matches being fair and wouldn't be driven to improve their decks. People who believe they can win are going to want the prizes.
Also I believe the f2p player base has enough bait on the hook already with pve so I don't believe the PvP base would expand enough to cover the loss of the tournament fees.
Also I elo just is a bad idea in general.

dogmod
08-25-2013, 10:37 PM
Are you saying casual players want a ranking system?

I am saying that casual players don't see the benefit in those points. I think that matching people of similar skill together would result in a more enjoyable experience on average for casual players, but the two points don't necessarily have to be linked.

zadies
08-25-2013, 10:43 PM
Actual I think the point he was making was that a ranking system only really benefits casuals.

Prodygi
08-25-2013, 10:51 PM
Actual I think the point he was making was that a ranking system only really benefits casuals.

He as in me? Or Dogmod?

zadies
08-25-2013, 10:58 PM
Pithe the person that posted the points that caused the question.

Chark
08-26-2013, 01:16 PM
Just to clarify my position:

I believe that casual players (players who are not competitive) do not care to see where they are in relation to other players in skill level. I do believe there is great value for them to get what they perceive is a fair matchup when they want to play a game. This reduces some of the anxiety of playing against unknown opponents. This can be accomplished with a match making system with a hidden rating.

Competitive players benefit very little from an exposed rating (see my previous post). I think the negative value an exposed rating would promote outweigh the positive value.

As a side note, I believe that it is extremely important to expose ratings (probably in terms of standings: "this is the 3rd best player in the world", etc.) for pro players. I define pro players as X% of the players at the top of any implemented tournament system. There is huge value in building stars for viewership and aspiration.

dogmod
08-26-2013, 01:29 PM
Just to clarify my position:

I believe that casual players (players who are not competitive) do not care to see where they are in relation to other players in skill level. I do believe there is great value for them to get what they perceive is a fair matchup when they want to play a game. This reduces some of the anxiety of playing against unknown opponents. This can be accomplished with a match making system with a hidden rating.

Competitive players benefit very little from an exposed rating (see my previous post). I think the negative value an exposed rating would promote outweigh the positive value.

As a side note, I believe that it is extremely important to expose ratings (probably in terms of standings: "this is the 3rd best player in the world", etc.) for pro players. I define pro players as X% of the players at the top of any implemented tournament system. There is huge value in building stars for viewership and aspiration.

I agree with these statements fully. I am for a hidden rating for competitive play and an exposed rating for Pro play. It will be interesting how they define the difference between the two and where the transition point will be.

BigDog
08-26-2013, 02:51 PM
I just wanted to mention something that may have been addressed elsewhere in the thread. M:tG elo ranking had issues regarding players being disincentivized to play once they reached a certain rating. This wasn't necessarily a problem with ELO ranking in and of itself as ELO is not as fitting for a card game as it is for a deterministic game such as chess. It was more of an issue of the rewards being tied to your ELO ranking rather than the ranking itself. So if you had ELO ranking just as a personal performance metric(aka epeen) not linked to rewards at certain numbers, it would be probably be fine and still useful for identifying areas for self-improvement. Obviously, this doesn't solve the matchmaking feature being discussed, but i wouldn't mind seeing ELO rankings even if they didn't affect anything

Prodygi
08-26-2013, 05:41 PM
Just to clarify my position:

I believe that casual players (players who are not competitive) do not care to see where they are in relation to other players in skill level. I do believe there is great value for them to get what they perceive is a fair matchup when they want to play a game. This reduces some of the anxiety of playing against unknown opponents. This can be accomplished with a match making system with a hidden rating.

Competitive players benefit very little from an exposed rating (see my previous post). I think the negative value an exposed rating would promote outweigh the positive value.

As a side note, I believe that it is extremely important to expose ratings (probably in terms of standings: "this is the 3rd best player in the world", etc.) for pro players. I define pro players as X% of the players at the top of any implemented tournament system. There is huge value in building stars for viewership and aspiration.

Agree with point 1 and 3. On the fence with point 2 though.

Having a check box for each individual player to choose whether to show or hide their ratings would be one solution.

Shrennan
08-26-2013, 08:00 PM
This is just an idea but I noticed how well it worked with Halo, so I thought I may suggest it here. Bungie.net had impeccable rating and ranking systems on their website that went into great detail on different matches, kill/death ratios, other ratios, in-depth looks at your characters, and eventually videos of matches on the site. I wouldn't be against CZE implementing a similar system for the Hex main site. I still think all of these things should be in the game, but if there were certain things that were easier to establish on a site that players could go to for even deeper statistics then I definitely don't think that would be a bad thing. It would also make it so that casuals wouldn't be overwhelmed by statistics within the game, as I'm sure there can be loads and loads of ranking statistics that would be quite daunting to the average, non-competitive player.

Please note, I'm only talking about ranking statistics, not general stuff that would be more suited for the double-back feature - I'm not talking about tournament wins or other PVP stuff that should be on the double-back.

Ebynfel
08-27-2013, 10:54 AM
I personally think that the players who make names for themselves without an inherent ratings system deserve more respect than those grinding out ELO. If you KNOW a player is damned good with no numbers to back it up, that's saying something.

Just saying.

Falaris
08-27-2013, 11:32 AM
Sorry if this is a point that has been alluded to/made previously:

If Cryptozoic wants to appeal to casual pvp play, then a set up similar to planeswalker points (PWP) would work just fine. They would be fine with having a system that is skewed towards people who play the most often as opposed to people who win a higher percentage of their matches.

However, if Cryptozoic is interested in a ranking system which more accurately details who the better players are, then an ELO-based system is clearly superior. Yes, ELO is far from perfect in a game that has random variables in it such as cards drawn for both you and your opponent. However, it is still a better determining factor as to who is better than a system that says a player that is 100-75 is better than a player that is 50-0.

One of the major strikes against ELO is that players who attain a certain ranking level are incentivized to sit on that ranking in order to accrue benefits befitting their rating. Aggressive rating decay is a perfectly fine answer to that problem. Giving a ratings penalty if a player doesn't play a certain number of matches in a week should prevent people from gaining a benefit to sitting on their rating. Finally, for people concerned that they can't play nonstop, something like a limited 'Airplane mode' seems like it wouldn't be difficult to implement. Enabling Airplane mode could prevent all play on the account (to prevent people from abusing taking a break from PvP to focus on PvE). It would be for situations like vacation and whatnot where you're not going to be online for an extended period but don't want to be penalized for taking a vacation, having a long work week, etc.

zadies
08-27-2013, 12:45 PM
Actually pro players are more likely to play for the rewards other then prestige and the idea presented here is that elo is done outside the tournament play where they can actually win boosters.