Ira Fay
United States New Haven CT
:)
Fay Games logo

Hello all! The US Tak Associate (USTA) is organizing a playtest of solutions to the first player advantage (FPA) issue. Tak has a pretty significant advantage to the first player, which is fine if you're not playing too competitively, or if you play pairs of games alternating colors.
But, in a tournament situation or in a situation where you want to play only a single game, it could be useful to have a version of Tak that's more balanced for both players. So, if you'd like to check out the possible ideas and help out, please complete this very short form (3 questions) to sign up: https://docs.google.com/forms/d/e/1FAIpQLSfCkGEAwG9xBhX8PJF...
Also, in case you don't know about it, a great website to play tak online for free is http://playtak.com
Thanks!


Lee Valentine
United States Boston Massachusetts

First player moves. Second player can accept the move and continue play, or the second player can swap colors and accept the original move as their own. This tends to make the first move a slightly suboptimal one, otherwise the second player would just steal the first player advantage.
I'd be interested in reading and commenting, but I don't use playtak.com.
Lee


Tysen Streib
United States San Francisco California

I think more than 1 move before the swap is probably better for Tak. I think:
First player puts down 24 pieces anywhere. Second player decides which color to play.


Lee Valentine
United States Boston Massachusetts

Now that I have got my Tak Companion, another potential solution is bidding back and forth.
Randomly determine the first bidder, who may start by passing. If that player passes on their first bid, their opponent is deemed to have made a bid of 1, and bidding ends.
Otherwise, the first bidder may start the bidding at any positive integer. Each player bids in turn again and again, with each bid being greater than the last until someone passes, which ends the bidding. A player can jump the bid to any higher integer and need not bid just one more than their opponent. The high bidder chooses whether to play first or second.
At the end of the ensuing game, the high bidder must ultimately beat their opponent by at least as many points as the high bid, or the high bidder is deemed to have lost the game to their opponent.
Obviously, such a bidding system requires sufficient knowledge of the game to meaningfully bid. This is also a good way to handicap the game: the more experienced player can bid high enough to make a victory quite difficult for themselves.
Lee


Mitch Willis
United States Kathleen Georgia

ira212 wrote: Tak has a pretty significant advantage to the first player How significant? What's the percentage of first player wins on playtak.com (I'm assuming that site can track such things)?


Travis Dean
United States Escondido CA
The Golden Camel UberBadge is a mystery. Nobody knows where it came from or how it got there.
No one expects the Spanish penguin!

otha62 wrote: ira212 wrote: Tak has a pretty significant advantage to the first player How significant? What's the percentage of first player wins on playtak.com (I'm assuming that site can track such things)?
There's much more discussion of this on Reddit. I think this was mentioned as well, and I don't remember exactly what it was, but I think I remember reading it's comparable to Chess. Something like 55/45% for a first/second player win. I could be misremembering though, since I'm not following the subreddit too closely.


Jason Webster
United States Connecticut

ira212 wrote: Hello all! The US Tak Associate (USTA) is organizing a playtest of solutions to the first player advantage (FPA) issue. Tak has a pretty significant advantage to the first player, which is fine if you're not playing too competitively, or if you play pairs of games alternating colors. But, in a tournament situation or in a situation where you want to play only a single game, it could be useful to have a version of Tak that's more balanced for both players. So, if you'd like to check out the possible ideas and help out, please complete this very short form (3 questions) to sign up: https://docs.google.com/forms/d/e/1FAIpQLSfCkGEAwG9xBhX8PJF...Also, in case you don't know about it, a great website to play tak online for free is http://playtak.comThanks!
Why in a tournament situation would switching first player and playing two games not be suitable?


Lee Valentine
United States Boston Massachusetts

Jason,
It would literally tend to double the length of a tournament. Ideally, you could calculate the first player advantage or bid for an offsetting handicap. You want to play two games only where time is not a factor or where that advantage has not been calculated.
Lee


Tysen Streib
United States San Francisco California

Dnasearchr wrote: Why in a tournament situation would switching first player and playing two games not be suitable? Plus if the advantage is significant, the results of 2 games would just be 1 win for each for a large percentage of matches.


Bill Reed
United States Galloway Ohio

tysen wrote: Dnasearchr wrote: Why in a tournament situation would switching first player and playing two games not be suitable? Plus if the advantage is significant, the results of 2 games would just be 1 win for each for a large percentage of matches.
Why is that?


Russ Williams
Poland Wrocław Dolny Śląsk

wmreed wrote: tysen wrote: Dnasearchr wrote: Why in a tournament situation would switching first player and playing two games not be suitable? Plus if the advantage is significant, the results of 2 games would just be 1 win for each for a large percentage of matches. Why is that? If there is a strong first player advantage, then by definition the first player will win most often between players somewhat close in strength.
E.g. suppose that between equal players, the first player wins 3/4 of the time. Then the probability that in a pair of games switching sides the first player wins both games (so each player wins once) is 9/16, more than half of the time.


Bill Reed
United States Galloway Ohio

russ wrote: wmreed wrote: tysen wrote: Dnasearchr wrote: Why in a tournament situation would switching first player and playing two games not be suitable? Plus if the advantage is significant, the results of 2 games would just be 1 win for each for a large percentage of matches. Why is that? If there is a strong first player advantage, then by definition the first player will win most often between players somewhat close in strength. E.g. suppose that between equal players, the first player wins 3/4 of the time. Then the probability that in a pair of games switching sides the first player wins both games (so each player wins once) is 9/16, more than half of the time.
Okay.
If there was no FPA issue, what's the significant difference when two equal players play?
Also I didn't think the calculated difference was as high as 3/4 of the time. Does that have a significant difference when playing a pair of games?


Russ Williams
Poland Wrocław Dolny Śląsk

wmreed wrote: If there was no FPA issue, what's the significant difference when two equal players play? In a game with no significant FPA, when 2 equal players, play, then of course each player has about 1/2 probability of winning the game, so the probabilities of each result in a pair of games with switching sides will be
Alfred wins 2 = 1/4 Alfred and Bob each win 1 (whether both as starter or both as second player) = 1/2 Bob wins 2 = 1/4
Quote: Also I didn't think the calculated difference was as high as 3/4 of the time. Does that have a significant difference when playing a pair of games? I don't know the FPA in Tak. I gather that it's indeed less than 3/4; I just gave 3/4 to show the effect of a clearly "significant" FPA.
Upthread dolus said 55% first player wins, so if that's true, then probability that each win 1 would be P(first player wins both times) + P(second player wins both times) = .55*.55 + .45*.45 = .505 which admittedly doesn't sound like a terrible problem.


Travis Dean
United States Escondido CA
The Golden Camel UberBadge is a mystery. Nobody knows where it came from or how it got there.
No one expects the Spanish penguin!

russ wrote: Quote: Also I didn't think the calculated difference was as high as 3/4 of the time. Does that have a significant difference when playing a pair of games? I don't know the FPA in Tak. I gather that it's indeed less than 3/4; I just gave 3/4 to show the effect of a clearly "significant" FPA. Upthread dolus said 55% first player wins, so if that's true, then probability that each win 1 would be P(first player wins both times) + P(second player wins both times) = .55*.55 + .45*.45 = .505 which admittedly doesn't sound like a terrible problem.
Found a link.
https://www.reddit.com/r/Tak/comments/4jo6wf/first_move_adva...
Quote: As there is a database of all games played available on playtak.com, it's easy to see who wins more often, searching for all 5x5 games gives that white completes 9674 roads to blacks 7803 at the time of writing, so of the games that ended by way of a road, white scored approximately 55% to blacks 45%, very similar to in chess.
However, this thread is from 8 months ago. I don't know how those percentages have changed over time.


Lee Valentine
United States Boston Massachusetts

Tysen, Tak is a pointscoring game, so it is irrelevant that the players each win one game in a twogame match unless their scores are identical. Without identical scores, playing two games and swapping who goes first is 100% decisive if each player's match score is the sum of their points won in both games. If a firstplayer handicap can be calculated or bid for, then a similar effect can be arrived at in a single game match, much as it is in go.
Lee


Jason Webster
United States Connecticut

hudarklord wrote: Jason,
It would literally tend to double the length of a tournament. Ideally, you could calculate the first player advantage or bid for an offsetting handicap. You want to play two games only where time is not a factor or where that advantage has not been calculated.
Lee
And what is the average length of a game ?


Lee Valentine
United States Boston Massachusetts

Jason, I am not sure of an exact number, but the average length of a game would vary greatly depending on the board size and the skill of the players involved. It also probably depends on whether there is an enforced time limit. If I have more than thirty minutes to finish a game, then I might use up all my time to consider every move against a highly skilled opponent, particularly if a lot of money was on the line. If you made me play under a tight time limit, then things would move more quickly.
Wikipedia guesses 90 minutes for playing Tak in a tournament, but I am unclear whether that is for one game or a round of two games. Either way, 90 minutes is a long time to play if you have a lot of rounds of a tournament to get through.
How many games is reasonable to play in one day depends on the style of tournament, the number of players, etc.
What's your best guess for the play time of a single tournamentlevel 6 x 6 game? For a match?
Lee


Russ Williams
Poland Wrocław Dolny Śląsk

hudarklord wrote: Tysen, Tak is a pointscoring game, so it is irrelevant that the players each win one game in a twogame match unless their scores are identical. Without identical scores, playing two games and swapping who goes first is 100% decisive if each player's match score is the sum of their points won in both games. If a firstplayer handicap can be calculated or bid for, then a similar effect can be arrived at in a single game match, much as it is in go. But it's not clear that one should use points that way. E.g. in Go tournaments, I've never seen point difference used for determining anything: the only relevant info is who won, and it doesn't matter if they won by 1 point or 50 points. A large point difference (in Go at least; I cannot say for Tak) does not necessarily indicate that the winner played much better or is a much stronger player. It could be an inevitable result of the game and how it develops: e.g. if a very large capturing race occurs, and both players play similarly well, then one player is going to have a very large group killed. It doesn't mean that they played much worse than the opponent.
Similarly in Tak, if the game ends by a connection, e.g. on a 5x5 board, the winner apparently earns 25 points. If the game ends with no connection made, but by controlling spaces with flat stones, then the winner probably has closer to 15 points. Is a connection win necessarily worth that much more (for tournament purposes) than a point win? Should it be worth that much more? Did a player who wins by points with 15 flat stones really play 40% worse than a player who wins by connection? Why isn't a win simply a win?


Jason Webster
United States Connecticut

hudarklord wrote: Jason, I am not sure of an exact number, but the average length of a game would vary greatly depending on the board size and the skill of the players involved. It also probably depends on whether there is an enforced time limit. If I have more than thirty minutes to finish a game, then I might use up all my time to consider every move against a highly skilled opponent, particularly if a lot of money was on the line. If you made me play under a tight time limit, then things would move more quickly.
Wikipedia guesses 90 minutes for playing Tak in a tournament, but I am unclear whether that is for one game or a round of two games. Either way, 90 minutes is a long time to play if you have a lot of rounds of a tournament to get through.
How many games is reasonable to play in one day depends on the style of tournament, the number of players, etc.
What's your best guess for the play time of a single tournamentlevel 6 x 6 game? For a match?
Lee
I have never played on a 6x6 board. Only 4x4 and 5x5. So I couldn't tell you.
Are tournaments only going to be 6x6??? Courtly rules.
I figured since Tak is suppose to be compared to chess and go you would use similar time constraints used in their tournaments. It would seem that playing by two games is still the fairest way to determine the winner. Even the companion books says it is the preferable way to play.
Finally, don't you feel that a bidding system changes the nature of the game? It's not Tak anymore is is a variant of the game. It becomes Tournament Tak instead of Tak.
What have players suggested was the best way to fix the first player advantage problem in your poll?


Jason Webster
United States Connecticut

russ wrote: hudarklord wrote: Tysen, Tak is a pointscoring game, so it is irrelevant that the players each win one game in a twogame match unless their scores are identical. Without identical scores, playing two games and swapping who goes first is 100% decisive if each player's match score is the sum of their points won in both games. If a firstplayer handicap can be calculated or bid for, then a similar effect can be arrived at in a single game match, much as it is in go. But it's not clear that one should use points that way. E.g. in Go tournaments, I've never seen point difference used for determining anything: the only relevant info is who won, and it doesn't matter if they won by 1 point or 50 points. A large point difference (in Go at least; I cannot say for Tak) does not necessarily indicate that the winner played much better or is a much stronger player. It could be an inevitable result of the game and how it develops: e.g. if a very large capturing race occurs, and both players play similarly well, then one player is going to have a very large group killed. It doesn't mean that they played much worse than the opponent. Similarly in Tak, if the game ends by a connection, e.g. on a 5x5 board, the winner apparently earns 25 points. If the game ends with no connection made, but by controlling spaces with flat stones, then the winner probably has closer to 15 points. Is a connection win necessarily worth that much more (for tournament purposes) than a point win? Should it be worth that much more? Did a player who wins by points with 15 flat stones really play 40% worse than a player who wins by connection? Why isn't a win simply a win?
Russ,if playing on a 5x5 board, when you win you score 25 point plus the number of pieces that remain in your reserve. It doesn't matter if you do that with a road win or a flat win.


Russ Williams
Poland Wrocław Dolny Śląsk

Dnasearchr wrote: Russ,if playing on a 5x5 board, when you win you score 25 point plus the number of pieces that remain in your reserve. It doesn't matter if you do that with a road win or a flat win. Doh! I brainfarted on how the Tak scoring works. Thanks for the correction!
...
Tak rules wrote: Scoring with Points: A game can be scored based on the number of pieces left unplayed. A winner is awarded “the board” (one point for every space on the board, for example 25), and “the pieces,” which is the number of his own pieces that remained out of play. This score is usually settled in coins. This is considered a “commoner’s” rule and is not observed in more sophisticated circles. At court, a win is simply a win. Hmm, but shouldn't a serious tournament be considered a more sophisticated circle, where a win is simply a win?
Is winning on a sparse board more glorious and more deserving of tournament points than winning on a full board? Is a solid strategy which builds to a slower victory and involves players placing more pieces inherently less worthy?
E.g. if I win with only 5 of my pieces on the board (i.e. with the fewest possible of my pieces on 5x5) making a straight connection, that sounds more like my opponent was completely incompetent rather than that I was a Tak genius.


Jason Webster
United States Connecticut

Is Playtak ever going to be iPhone friendly? I rarely have computer access


Jason Webster
United States Connecticut

russ wrote: Dnasearchr wrote: Russ,if playing on a 5x5 board, when you win you score 25 point plus the number of pieces that remain in your reserve. It doesn't matter if you do that with a road win or a flat win. Doh! I brainfarted on how the Tak scoring works. Thanks for the correction! ... Tak rules wrote: Scoring with Points: A game can be scored based on the number of pieces left unplayed. A winner is awarded “the board” (one point for every space on the board, for example 25), and “the pieces,” which is the number of his own pieces that remained out of play. This score is usually settled in coins. This is considered a “commoner’s” rule and is not observed in more sophisticated circles. At court, a win is simply a win. Hmm, but shouldn't a serious tournament be considered a more sophisticated circle, where a win is simply a win? :) Is winning on a sparse board more glorious and more deserving of tournament points than winning on a full board? Is a solid strategy which builds to a slower victory and involves players placing more pieces inherently less worthy? E.g. if I win with only 5 of my pieces on the board (i.e. with the fewest possible of my pieces on 5x5) making a straight connection, that sounds more like my opponent was completely incompetent rather than that I was a Tak genius. :)
Your last statement is true. So while that score would not show you were a Tak genius it definitely would show which player should advance in a tournament setting.


Russ Williams
Poland Wrocław Dolny Śląsk

Dnasearchr wrote: Your last statement is true. So while that score would not show you were a Tak genius it definitely would show which player should advance in a tournament setting. Certainly between those 2 particular players. But I would equally be selected to advance versus my incompetent opponent if we merely counted number of wins.
But if I win with only 5 stones placed against my clearly incompetent player, and meanwhile Bob wins his game with 13 stones placed, why should I be awarded more tournament points than Bob?


Jason Webster
United States Connecticut

russ wrote: Dnasearchr wrote: Your last statement is true. So while that score would not show you were a Tak genius it definitely would show which player should advance in a tournament setting. Certainly between those 2 particular players. But I would equally be selected to advance versus my incompetent opponent if we merely counted number of wins. But if I win with only 5 stones placed against my clearly incompetent player, and meanwhile Bob wins his game with 13 stones placed, why should I be awarded more tournament points than Bob?
Ah. I have never played in any game tournaments. I figured it was a ladder tournamebtbof some sort where the two finalist met in a glouriously contested final game. My lack of understanding tournament types is showing.
Hmmm. How was the first Tak tournament conducted. Was winner done by points played over a number of games or did two finalists play head to head for the championship?



