A great database for the community

[Tyr Anasazi]

I guess that's the safer approach but then it would discard games where people quit because they know they will lose (happens very often on QM). Not sure how that might bias the data.



Almost any player who keeps his own stats (usually players averaging 300 rating) would tell you that FTA is more than 3% by their own stats. Evidently this discrepancy with the stat given by Kyle (which is obviously true; no one thinks he's lying) must have some explanation.

It's not that players are so rubbish that their chance of winning is 50% (besides since people like Preybird have a higher than 50% win rate, weaker players would be expected to have less than 50% chance to win). It's that at lower rating, skill level plays a much larger role in determining win rate compared to match-up/going first/other factors. Therefore, going 1st or playing a specific hero has a weaker relationship with winning compared to at higher ratings.

I leave you again with the same data from chess:
http://en.wikipedia.org/wiki/First-m...ntage_in_chess


The wiki has the actual citation if you doubt its credibility.

I do believe, just in theory, a game such as SE where luck is involved tends to have a smaller FTA than a full deterministic game such as chess.

The reason is simple. In chess, once an opening that gives significant advantage to White is discovered, it can be reliably repeated ever since.

But in SE, even if you know a God hand will make you win 99.9% of the time, the chance of drawing that God hand is nowhere near 99.9%. So it cannot be reliably repeated game after game.

[pyrogene]

I do believe, just in theory, a game such as SE where luck is involved tends to have a smaller FTA than a full deterministic game such as chess.

You can only say the SE with randomness has less FTA than SE without randomness (i.e. a game where you can pick which card you draw). You can't really compare it with Chess that way since they are so fundamentally different. E.g. Playing rush Eladwen is more random and luck based than chess but it will show significantly higher FTA compared to Chess.

[Tyr Anasazi]

You can only say the SE with randomness has less FTA than SE without randomness (i.e. a game where you can pick which card you draw). You can't really compare it with Chess that way since they are so fundamentally different. E.g. Playing rush Eladwen is more random and luck based than chess but it will show significantly higher FTA compared to Chess.

Okay. So Chess may not be the best example since so far Chess is still not solvable due to its complexity.

But for all deterministic games, including Chess, once it is solved, it is either white always wins (100% FTA) or black can always draw (0% FTA). One example of such solved game is Gomoku, where player going 1st always wins with traditional rules.

But games such as Poker or SE, even after solved, will be something like 60% of the time FTA win you the game due to inherent randomness. It will never to as extreme as 100% or 0%. And that's the point I tried to make.

Back to your Elad rush example, although unlikely, it can happen that opponent's hero has been reduced to 1 HP, however all you 12 burns are at the bottom of your deck and you lose as a result. This will almost never happen in Chess.

[a player]

I wonder why some matches end in Draw!, even in Meltdown format, and why do some games not leave a record in one's history. It could be interesting to review a game that suddenly ended on turn four with "too short to gain experience", or to be able to review what opposing heroes are most likely to drop out in the first three turns.

So far I haven't been tracking my stats, simply because it seems like too much work, but that is precisely the stuff that computers excel at. I understand the developers not wanting all information to be leaked to everyone, but I still wish that more of our personal stats were tracked and made available to us, individually.

I'd also like to see rudimentary analysis of our decks in the deckbuilder: quantity vs. cost curve, allies/abilities/weapons percentages, etc.

[pyrogene]

Okay. So Chess may not be the best example since so far Chess is still not solvable due to its complexity.

But for all deterministic games, including Chess, once it is solved, it is either white always wins (100% FTA) or black can always draw (0% FTA). One example of such solved game is Gomoku, where player going 1st always wins with traditional rules.

But games such as Poker or SE, even after solved, will be something like 60% of the time FTA win you the game due to inherent randomness. It will never to as extreme as 100% or 0%. And that's the point I tried to make.

Back to your Elad rush example, although unlikely, it can happen that opponent's hero has been reduced to 1 HP, however all you 12 burns are at the bottom of your deck and you lose as a result. This will almost never happen in Chess.

That was precisely what I was saying when I said SE with randomness would have less FTA than SE without. But since SE is not played without randomness and Chess/"deterministic" games are not played with randomness, you can't really compare between them.

SE can be luck-based and still have significantly more FTA than a non-random game.

[plo]

Kyle: Now that DP has finally been released would it be possible to reveal the CotC data? I suppose it would not cause trouble because that information is not true anymore. I'm just asking out of curiosity.