|Dominion Woodcutter Analysis|
|I’m using logged 2-player games from ceviri’s woodcutter.|
Before Renaissance, those are games from January 2018 to 14 September 2018 in which one opponent had at least mu=1.9 or is one of the named players with detailed statistics.
Since Renaissance, I'm using all games in which the better player had at least mu=1.8 or the worse player had at least mu=1.3.
I’m excluding games with more than 2 landscapes and games that ended before turn 3 of player 1.
|Some caveats:||Last decision is not logged. This could be ending your buy phase, but it could also be buying the last Province if you can’t take any decision afterwards.|
Band of Misfits and Overlord had log issues before Renaissance. Only games since Renaissance release are included. Similarly, a few other games can’t be loaded. If some fancy play was responsible for that, it would be systematically excluded from the statistics.
Sample size: when it comes to statistics for a particular card and player, there are fewer games making the estimates not very precise. More generally, if you look at over 300 cards, you’re expected to find a bunch of outliers because of sampling errors.
|IMAGES WITH SUMMARY STATS FOR ALL CARDS||explanatory video by aku chi|
|IMAGES WITH SUMMARY STATS ONLY FOR CARDS FROM RENAISSANCE|
|players with 100 logged games||available sheets with detailed statistics|
|PerryGreen||1371||For each player there are 8 tabs:|
|Mercury444F||1358||buys,gains,trashes: shows the percent of games each “card” is bought, gained and trashed by the considered player and their opponent as well as the average number of those buys, gains and trashes.|
|nakabisya36||1337||b,g,t detail: has the distribution of those buys, gains and trashes for the player of interest.|
|Gherald||1306||gain 1st: shows how likely a card is the first to be bought (or gained / trashed) once, twice, three times,...|
|gloures||1228||gain 1st Qvist: shows the same as gain 1st but within the cost categories in the Qvist ranking (e.g. first $5 cost to be bought once or twice.|
|LimeTime||1158||boards: shows some averages for games that had a particular card in the kingdom. First the number of turns (all turns by both players including extra turns are added). Then the total number of cards bought, gained and trashed by each player. The last columns have the win-loss statistic for the first player and for the player of interest on those boards. The first column of outperform shows how much the performance was above expectation according to the players’ skill mu. For example, if a player was expected to win 70% of the games, but they won 75% it will show up as +5%. The column with +/- has the standard error of this estimate and indicates how serious you should take the outperformance. Keep in mind that in this list there should be 15-20 cards for which the outperformance is off by more than twice the value in +/-.|
|vsiewnar||940||outperformance: performance of that player on those boards|
|wharf_rat||919||opponents: has the win-loss statistics for the opponents of specific players.|
|sangatsu||896||loggedgames: has the game numbers used for that player linked to woodcutter and the two players, their skills (mu), the winner, and the cards in the kingdom.|
|Obi Wan Bonogi||452|