|
|
|
Statistics of Abstract Strategy Games
|
|
|
This is a fledgling list of abstract strategy games, which are played over the web or in a competitive manner. It will give game statistics that are crucial for determining the quality of a game. The games have perfect information (no hidden information whatsoever) and no random elements (initial position must also be the same in all games). Luck may play no role (the players are solely responsible). I believe that, on the average, a game should take at least 40 turns (or half-moves) to have enough "depth", but should not be longer than 100 turns , otherwise I would call the game "lengthy". A game should be "well-balanced", too. A perfect balance would be if either player has a 50% chance to win (it is supposed that both players are equally strong). A game is not well balanced, if one player has a chance of more than 55% to win. The percentage of draws should not be larger than 10% among amateurs, otherwise I call the game "drawish", and no further analysis will be conducted. If a game is "recommended", I think that it is suited for tournament play. "Not recommended" means: "I wouldn't recommend the game to be used in tournaments." The game can still be beautiful and may work fine in casual play. If a game differs by a factor of 3 from what I defined above (e.g. the percentage of draws among amateurs is larger than 30% or one player has a chance to win the game of more than 65%), I call the game "flawed". Games are only added when I can find at least a sample of 50 games. I will add 1-2 games each month. This is an attempt to provide objective information on the quality of games. I find it sad that some of my favorites ended among the "not recommended" games, however, I find it even sadder that some people think they have to use tags to spread misinformation because their favorite game lacks the quality an abstract strategy game should have.
|
|
|
|
|
Posted On: 2007-09-28 03:16:07
Edited On: 2007-10-24 08:56:49
|
|
1.
Abande [Average Rating:7.39 Unranked]
|
|
|
|
|
111 games played at Super Duper Games were analyzed (hexagonal board only; abandoned games omitted).
The average length of a game was 53 turns. The longest game had 63 turns.
50% (53) of the games, which didn't end in a draw, were won by the second player, 50% (52) by the first player.
5.5% (6) of all games were drawn.
Abande is WELL BALANCED and has a REASONABLE LENGTH.
The game is RECOMMENDED. 
|
|
|
|
0 Comments
|
Posted On: 2007-09-28 04:12:59
Edited on: 2007-09-29 11:23:14
|
|
|
|
|
58 games played at Super Duper Games were analyzed (abandoned games omitted).
The average length of a game was 28 turns. The longest game had 65 turns.
57% (33) of the games were won by the first player, 43% (25) by the second player.
Archimedes is NOT WELL BALANCED and is TOO SHALLOW.
The game is NOT RECOMMENDED.  |
|
|
|
0 Comments
|
Posted On: 2007-09-28 05:31:28
Edited on: 2007-09-28 07:36:09
|
|
|
|
|
70 games played at Super Duper Games were analyzed (abandoned games, square board and "open" variant omitted).
The average length of a game was 48 turns. The longest game had 169 turns.
59% (41) of the games were won by the first player, 41% (29) by the second player.
Attangle is NOT WELL BALANCED.
The game is NOT RECOMMENDED.
Note:
I divided the above sample in five subsamples of equal size (ie. 14 games each) and counted the number of wins for the first player. In the next step, I calculated the empirical variance for the game's balance. The standard error turned out to be less than +/- 1.6% despite the small subsamples. In fact the standard error must be even smaller as the sample that was used to calculate the balance of the game was much larger. |
|
|
|
0 Comments
|
Posted On: 2007-09-28 07:35:17
Edited on: 2007-10-02 03:12:10
|
|
4.
Blam! [Average Rating:6.32 Unranked]
|
|
|
|
|
83 games played at Super Duper Games were analyzed (no "overloaded" variant; abandoned games omitted).
The average length of a game was 52 turns. The longest game had 96 turns.
59% (48) of the games, which didn't end in a draw, were won by the second player, 41% (33) by the first player.
2.5% (2) of all games were drawn.
Blam! is NOT WELL BALANCED.
The game is NOT RECOMMENDED. |
|
|
|
0 Comments
|
Posted On: 2007-10-01 06:15:26
Edited on: 2007-10-01 06:16:13
|
|
|
|
|
85 games played at Super Duper Games were analyzed (standard board (5x5) only; abandoned games omitted).
The average length of a game was 131 turns. The longest game had 150 turns.
54% (46) of the games were won by the first player, 46% (39) by the second player.
Cephalopod is WELL BALANCED, but the game is rather LENGTHY.
It is ONLY RECOMMENDED, if you have LOTS OF TIME.  |
|
|
|
0 Comments
|
Posted On: 2007-10-02 03:05:55
Edited on: 2007-10-03 04:25:17
|
|
|
|
|
96 games played at Super Duper Games were analyzed (abandoned games omitted).
The average length of a game was 38 turns. The longest game had 92 turns.
52% (50) of the games were won by the second player, 48% (46) by the first player.
Chase is TOO SHALLOW.
The game is NOT RECOMMENDED. |
|
|
|
0 Comments
|
Posted On: 2007-10-03 04:19:22
|
|
|
|
|
54 games played at Super Duper Games were analyzed (abandoned games omitted).
The average length of a game was 70 turns. The longest games had 91 turns.
50% (27) of the games were won by the first player, 50% (27) by the second player.
Diffusion is WELL BALANCED and has a REASONABLE LENGTH.
The game is RECOMMENDED. |
|
|
|
0 Comments
|
Posted On: 2007-09-29 11:18:44
|
|
|
|
|
97 games played at Super Duper Games were analyzed (abandoned games omitted).
The average length of a game was 22 turns. The longest game had 65 turns.
53% (47) of the games, which didn't end in a draw, were won by the second player, 47% (42) by the first player.
8% (8) of all games were drawn.
Fanorona is TOO SHALLOW.
The game is NOT RECOMMENDED.
Note:
I divided the above sample in four subsamples of equal size (ie. 22 games, omitting the last one and all drawn games) and counted the number of wins for the first player. In the next step, I calculated the empirical variance of the game's balance. The standard error turned out to be less than +/- 3.0% despite the small subsamples. In fact the standard error must be even smaller as the sample that was used to calculate the balance of the game was much larger. |
|
|
|
0 Comments
|
Posted On: 2007-09-29 03:37:47
Edited on: 2007-10-02 03:14:10
|
|
|
|
|
59 games played at Super Duper Games were analyzed (abandoned games omitted).
The average length of a game was 89 turns. The longest game had 120 turns.
51% (30) of the games were won by the first player, 49% (29) by the second player.
Impasse is WELL BALANCED and has a REASONABLE LENGTH.
The game is RECOMMENDED. |
|
|
|
0 Comments
|
Posted On: 2007-09-28 14:19:48
|
|
|
|
|
167 games found on Boardspace.net were analyzed. Games, in which guests or bots participated, were excluded. Also, no games were counted, in which people played against themselves or an irregular board was used.
The average length of a game was 63 turns. The longest game had 126 turns.
55% (92) of the games were won by the second player, 45% (75) by the first player. No draws could occur as Soucie's original rule was implemented (simultaneous connection counts as a win for the moving player).
Lines of Action is WELL BALANCED and has a REASONABLE LENGTH.
The game is RECOMMENDED. |
|
|
|
0 Comments
|
Posted On: 2007-09-30 01:39:07
|
|
|
|
|
52 games played at Super Duper Games were analyzed (standard rules only (no "weak force" variant); abandoned games omitted).
The average length of a game was 33 turns (but see note below). The longest game had 342 turns (no mistake!).
52% (27) of the games were won by the first player, 48% (25) by the second player.
Magneton is WELL BALANCED. It's too early to determine its length.
The game is STILL INVESTIGATED. 
Note:
Many games ended because one player accidentally achieved the winning position for his/her opponent. Later a variation was implemented, in which a player is warned before he can do such a mistake. This implementation was used only a few times at SDG. It resulted in two very long games, one had 297 turns, the other one took 342 turns. I believe that careful play between experienced players will lead to lengthy games and, perhaps, to games that will go on forever. |
|
|
|
0 Comments
|
Posted On: 2007-10-01 02:10:30
Edited on: 2007-10-02 01:58:21
|
|
|
|
|
90 games played at Super Duper Games were analyzed (two-player games only; goal: 12 points; abandoned games and games with different goals (10 points, 15 points) were omitted).
The average length of a game was 44 turns. The longest game had 87 turns.
50% (45) of the games were won by the first player, 50% (45) by the second player.
Pikemen is WELL BALANCED and DEEP.
The game is RECOMMENDED. |
|
|
|
0 Comments
|
Posted On: 2007-09-28 13:09:59
|
|
|
|
|
63 games found on Boardspace.net were analyzed (abandoned games omitted). Games, in which guests or bots participated, were excluded.
The average length of a game was 52 turns. The longest game had 154 turns.
63% (40) of the games were won by the second player, 37% (23) by the first player.
Pünct is NOT WELL BALANCED.
The game is NOT RECOMMENDED. |
|
|
|
0 Comments
|
Posted On: 2007-10-02 04:49:29
Edited on: 2007-10-02 04:51:14
|
|
|
|
|
53 games played at Super Duper Games were analyzed (abandoned games were omitted).
The average length of a game was 26 turns. The longest game had 48 turns.
52% (25) of the games were won by the second player, 48% (23) by the first player.
Quadrature is TOO SHALLOW.
The game is NOT RECOMMENDED. |
|
|
|
0 Comments
|
Posted On: 2007-09-28 08:46:29
Edited on: 2007-09-28 08:46:58
|
|
| |
