PDA

View Full Version : Stats, Stats, and more Stats



redyetiking
06-03-2013, 03:25 AM
Please give me as much stats about how I play as you can. I want to know what card I play most. how many times I play a card. What is my win/lose and any other stats that will help me be a better player.

Just give me the data.

wallofomens
06-03-2013, 03:59 AM
I don't think that anything you mentioned would make you a better player. But yeah, data is always nice to have.

Rapkannibale
06-03-2013, 05:07 AM
Best thing will be to record your matches and re-watch them afterwards. Stats are nice to have and can help in certain situations I guess.

Schweinebub
06-03-2013, 06:49 AM
I'd like to see a killing blow stat on each card. How many of my games did I win by burning my opponent? What's the percentage of games this card was the final blow of the total number of game the card was actually in my deck, broken down by PVE and PVE.
Another cool stat would be the average damage a card does each time you play it.

There are so many things they could track and I want them all!

jai151
06-03-2013, 06:54 AM
With a purely digital cardspace, they could do some crazy analytics. The question is if it's worthwhile to do so.

Mr.Funsocks
06-03-2013, 08:26 AM
Any scientist can tell you: Stats MEAN something, when you know how to read 'em. I would love for there to be crazy analytics like that, and it's entirely within reason for the game.

Tyrfang
06-03-2013, 08:31 AM
Any scientist should also realize that statistics can be used to lie just or be misinterpreted much more easily than correctly interpreted...

Throw enough numbers around, and someone's going to bitch about something.
"X is OP because 60% of players play it in their deck!" - except x is in a starter, so everyone has them
"Y is UP because only 2% of players play it in their deck!" - except y decks have a 60% win rate
"Z is OP because it has a 60% win rate!" - except Z is in 60% of decks.

Keep it simple.

Mr.Funsocks
06-03-2013, 08:34 AM
Any scientist should also realize that statistics can be used to lie just or be misinterpreted much more easily than correctly interpreted...

Throw enough numbers around, and someone's going to bitch about something.
"X is OP because 30% of players play it in their deck!" - except x is in a starter, so everyone has them
"Y is UP because only 2% of players play it in their deck!" - except y decks have a 60% win rate
"Z is OP because it has a 60% win rate!" - except Z is in 60% of decks.

Keep it simple.

Well duh :-P

But we weren't seeking global stats (though they could be neat...), we were talking personal metrics.

Punk
06-03-2013, 08:34 AM
Best thing will be to record your matches and re-watch them afterwards. Stats are nice to have and can help in certain situations I guess.

I can't agree more.

This was a tactic I stole from a friend many years ago. He would re-watch all of his online poker tournaments to analyze where he made mistakes, where he could improve, etc. I will go back and watch Drafts where I didn't win (and even some that I do) to do the same thing. I have become a much better player since I have started doing this. One very difficult thing to overcome is criticizing yourself and your actions. Once you can realize your own mistakes and learn from them, you will become a very good player.

RobHaven
06-03-2013, 08:34 AM
Resources are OP - 100% of all winning decks contain them.

ramseytheory
06-03-2013, 08:36 AM
I'd certainly find a win-loss ratio - preferably with the ability to adjust the timeframe and game types considered - very useful. Among other things it would give me a good indication of whether I should be in Swiss, casual or competitive draft. Stats associated with each of my decks would be nice as well.

Milamber
06-03-2013, 08:39 AM
+1 to cool metrics. there was a CZE post somewhere listing things like , max damage, total damage, kills etc.

Someone with a better memory and good "link fu"
... anyone?!

RobHaven
06-03-2013, 08:42 AM
Stats associated with each of my decks would be nice as well.

The issue with stats comes when variables aren't accounted for. Reading stats on a deck level would be worthless unless framed in the proper context; what if you change out cards X, Y, and Z for cards A, B, and C? It's not the same deck anymore. And if your sample size is too small, the composition of the decks you're playing against will have an enormous impact on how your deck's stats appear.

tgm0112
06-03-2013, 08:48 AM
Any scientist can tell you: Stats MEAN something, when you know how to read 'em. I would love for there to be crazy analytics like that, and it's entirely within reason for the game.

Actually any mathematician (heyo!) will tell you that is not the case. Some statistics are devoid of meaning. There's such a thing as signal to noise ratio.

Not all data is good data and fitting for noise is one of the worst things you can do.


EXAMPLE

When card Alpha is in a deck it has people lose 95% of the time. This does not imply on its own that card alpha shouldn't be used. It might mean that card alpha has a specific niche that is being ill-employed by inexperienced masses (a HUGE source of noise).

Mr.Funsocks
06-03-2013, 08:50 AM
Actually any mathematician (heyo!) will tell you that is not the case. Some statistics are devoid of meaning. There's such a thing as signal to noise ratio.

Not all data is good data and fitting for noise is one of the worst things you can do.

Hence why I said "when you know how to read them" :-P

Numbers don't lie, some people just don't speak Number.

Tyrfang
06-03-2013, 08:53 AM
Numbers don't lie.

Statistics, however, can indeed "lie" with the proper interpretation.

EDIT: Anyway, this is getting off topic.

I would support adding more card stats at the player-level, but I would avoid anything like aggregated win/loss stats for cards/combos/decks.

Punk
06-03-2013, 08:58 AM
Resources are OP - 100% of all winning decks contain them.

100% of all winning decks draw cards.

I think we're on to something..

Gwaer
06-03-2013, 09:08 AM
100% of all losing decks have at least one threshhold though =/

tgm0112
06-03-2013, 09:09 AM
@Mr.Funsocks

Tyrfang said it rather well. Numbers don't lie, but stats can be unhelpful or even deceitful when the wrong context is applied. You can have a distribution with an unbounded mean, but any time you poll from this distribution you can calculate an average, which is meaningless.

larryhl
06-03-2013, 09:13 AM
100% of all losing decks have at least one threshhold though =/

You could technically keep a no resource hand and just not draw resources for the loss.

Mr.Funsocks
06-03-2013, 09:24 AM
@Mr.Funsocks

Tyrfang said it rather well. Numbers don't lie, but stats can be unhelpful or even deceitful when the wrong context is applied. You can have a distribution with an unbounded mean, but any time you poll from this distribution you can calculate an average, which is meaningless.

Yes, if you're dumb and don't understand statistics, they're misleading :-P I've said that many times. And admittedly, most people ARE dumb and don't understand statistics.

tgm0112
06-03-2013, 09:29 AM
Yes, if you're dumb and don't understand statistics, they're misleading :-P I've said that many times. And admittedly, most people ARE dumb and don't understand statistics.

You misunderstand me. It's a matter of unknown unknowns. It's not that one is dumb if they're misled; they might simply not have enough information to make the right conclusion and be incapable of ever doing so (and not for lack of intellect). Moreover, if the signal-to-noise ratio is bad enough, as it may be with a game that people will do a lot of experimenting in, the statistics might actually support a false conclusion.

This is symptomatic of a lot of dynamical systems (e.g. earthquakes, weather). This type of TCG is no different. A small change in initial conditions, your opening hand in this case, has massive consequences for the final outcome.

larryhl
06-03-2013, 09:30 AM
Knowing when to mulligan is definitely a skill in TCGs. But that's off-topic.

I think the issue is that there will be a lot of stats, but without the right "tags" applied to them or knowing the circumstances that something happened, they really aren't meaningful.