Essay:Exploring Draws as a Replacement for Going to Time Rules

From Archon Arcana - The KeyForge Wiki


While there are many things KeyForge does well, the rules for Going to Time is not one of them; a game that goes to time in an Organized Play event can be very frustrating for both players involved, not to mention the judges who have to preside over the procedure. The Going to Time rules create two very distinct problems. The first issue I will touch on briefly here, but the rest of this essay will focus on the second.

Issue One

An Archon Solo match must complete in 35 minutes, and the current going to time rules mean that if a player can get to a point where they have an advantage in Keys or Æmber, they have no motivation to yield time to their opponent. Multiple people have suggested that implementing chess clocks would be a good solution to this problem, and I don't disagree that it is a valid solution. KeyForge at its heart (at least currently) is an asynchronous game. While one player is taking their turn, the other player has no options or obligations to react to anything being done. Implementing chess clocks would solve this problem by altering the win conditions; the winner of the game is the person who forges three keys or the person who did not run out of time. Chess clocks certainly are a valid potential solution to this issue. They could also create a better handicapping mechanism for KeyForge than chains currently provide; players can adjust with more or less time based on their experience and skills with a deck. However, chess clocks also create a barrier to entry to the game that would require mitigation for them to be widely accepted.

Issue Two

The Going to Time rules do not often accurately reflect which player performed well enough to be credited with a tournament win for the round. As mentioned above, KeyForge's asynchronous game structure means that a player can only respond to the game state during their turn. This creates a feel to the game that is often described by players as "swingy", meaning that after a turn or more of doing very little, a player is in a place where they can perform multiple game actions and make significant strides towards a victory. In casual (untimed) play, this is what makes KeyForge exciting, allowing for explosive turns and come from behind victories. In a timed setting, the going to time rules essentially slam the brakes on the game and then assign the tournament point for the round to which ever player came closest to winning. With the swingy nature of the game, this often leaves one or both players feeling disappointed, the winning player feels bad about taking a win from a 2 key to 2 key draw where they had slightly more Æmber, while the losing player is left scratching their head as to what they possibly could have done better (or worse, looking across the table at their opponent and wondering if their hesitation to play was due to careful consideration of their turn or something more nefarious.

The Proposal

In an attempt to deal with Issue Two, I propose that KeyForge Organized Play implement a system where if a match is not completed within the prescribed time limits, both players receive a draw which is worth one quarter of a tournament point. This proposal acknowledges KeyForge's asynchronous game structure creates a scenario where, if a game is not completed, a winner cannot be accurately determined. Since no winner can be determined, both players receive partial credit for a win. However, this partial credit should be set low enough so as not to encourage "gaming the system" to advance in a tournament round.

Experiment One

Based on KeyForge Tournament Structures, if a tournament has 77-148 players, the players will play 6 rounds of Swiss, resulting in a top cut of 16 players. Since this size tournament covers a large part of previous Vault Tour player counts, I used a player count of 78 players. Utilizing FFG's TOME software (which utilizes the same backed as GEM, FFG's Organized Play software), multiple tournaments were built and compared to a "control tournament" in which no games went to time. In order to facilitate the tournament software, a win is give 4 points, while a loss was given 0 points. This allowed for draws to be selected at either a quarter or a half TP. All tournaments used the same 6 rounds of 39 games to be "played", for a total of 234 games.

These experimental tournaments assume that Player 1 is an exceptionally good player, and that if the game is completed within the tournaments time limit, Player 1 will always win. Since the Going to Time rules do not necessarily reflect which player performed well enough to be credited with a tournament win for the round, Player 1 can lose when a match goes to time. Based on this assumption, Player 1 should finish at the top of the top 16 cut.

Control Tournament

In this tournament, all games were considered to have finished on time, so no draws were required. Players are identified by number (Player 1, Player 2, etc.) and the player with the lowest identifying number (not rank) was selected to win the match. The scenario would be unchanged if Player 1 went to time in round 1 and was given the tournament points for the win.

RANK    PLAYER          PTS     SOS     ESOS
1.	Player	1	24	2.66667	2.46296
2.	Player	2	24	2.44444	2.48148
3.	Player	5	20	2.66667	2.25926
4.	Player	10	20	2.55556	2.33333
5.	Player	4	20	2.44444	2.34722
6.	Player	13	20	2.22222	2.26852
7.	Player	12	20	2.22222	2.45833
8.	Player	14	20	2.22222	2.38889
9.	Player	11	20	2.00000	2.18519
10.	Player	7	16	3.00000	2.27778
11.	Player	3	16	3.00000	2.14815
12.	Player	16	16	2.66667	2.29167
13.	Player	27	16	2.44444	2.20370
14.	Player	19	16	2.44444	2.31481
15.	Player	8	16	2.44444	2.19907
16.	Player	18	16	2.33333	2.25926

Round One Loss

In this tournament, Player 1 and their opponent go to time in the first round, but Player 1 does not receive the tournament points from going to time. In this case, Player 1 still makes the top cut, but has dropped to 8th place.

RANK    PLAYER          PTS     SOS     ESOS
1.	Player	2	24	2.88889	2.42593
2.	Player	3	20	2.88889	2.31481
3.	Player	5	20	2.88889	2.42593
4.	Player	4	20	2.77778	2.51852
5.	Player	8	20	2.33333	2.53704
6.	Player	6	20	2.22222	2.31481
7.	Player	13	20	2.22222	2.27778
8.	Player	1	20	2.00000	2.22222
9.	Player	9	20	1.88889	2.20370
10.	Player	7	16	2.77778	2.42593
11.	Player	14	16	2.77778	2.33333
12.	Player	11	16	2.66667	2.18519
13.	Player	10	16	2.66667	1.98148
14.	Player	20	16	2.44444	2.33333
15.	Player	15	16	2.44444	2.14815
16.	Player	17	16	2.33333	2.22222

Round One Draw

In this tournament, Player 1 and their opponent go to time in the first round, and they each receive a partial tournament point for the round (.25 points in the example below). Player 1 finished second in the tournament, and there would be no difference if Player 1 were awarded a half point or a quarter point.

RANK    PLAYER          PTS     SOS     ESOS
1.	Player	4	24	2.66667	2.65278
2.	Player	1	21	2.13889	2.15741
3.	Player	5	20	2.88889	2.43519
4.	Player	3	20	2.77778	2.40278
5.	Player	2	20	2.36111	2.26852
6.	Player	6	20	2.33333	2.39352
7.	Player	18	20	2.22222	2.35185
8.	Player	7	20	2.11111	2.31481
9.	Player	16	20	2.00000	2.17130
10.	Player	12	16	2.80556	2.35648
11.	Player	11	16	2.69444	2.22685
12.	Player	10	16	2.66667	2.15278
13.	Player	15	16	2.58333	2.32407
14.	Player	9	16	2.55556	2.26389
15.	Player	19	16	2.55556	2.22222
16.	Player	13	16	2.55556	2.20833


Since Player 1 is expected to finish at the top of the tournament, we see that using a system that will consistently give Player 1 some points when they are involved in a match that goes to time means the result of the tournament will more closely resemble the results of the Control tournament than a situation where Player 1 receives a loss from going to time.

Experiment Two

One concern of implementing draws is whether or not a player could consistently try to draw games and still make the top cut. In this experiment, Player 1 is no longer an undefeatable player, but rather is a player who strives to go to time, because they have found that while they cannot out and out defeat an opponent, they can get to an advantageous position a certain portion of the time, getting enough wins to make the top cut. For example, if Player 1 goes to time and gets credited with a win for the first 4 rounds, they can finish in the top cut as high as 9th:

RANK    PLAYER          PTS     SOS     ESOS
1.	Player	4	24	2.44444	2.38889
2.	Player	2	24	2.44444	2.37037
3.	Player	6	20	2.88889	2.42593
4.	Player	3	20	2.55556	2.38889
5.	Player	5	20	2.44444	2.38889
6.	Player	10	20	2.44444	2.38889
7.	Player	8	20	2.44444	2.24074
8.	Player	23	20	2.33333	2.29630
9.	Player	1	16	2.88889	2.05556
10.	Player	7	16	2.77778	2.37037
11.	Player	20	16	2.55556	2.20370
12.	Player	15	16	2.55556	2.14815
13.	Player	13	16	2.55556	2.12963
14.	Player	16	16	2.44444	2.29630
15.	Player	11	16	2.44444	2.25926
16.	Player	14	16	2.44444	2.16667

Since a win when going to time is counted for the same value as a forged keys win, this creates a situation in which a Player now has two paths to ways to win a tournament round; they can either forge three keys, or they can be in an advantageous position when time is called.

Partial Point Scoring

Allowing for a draw to replace the going to time rules will do a lot to hamper Player 1, since they can no longer consistently rely on scoring a full tournament point with their stalling deck. The next question is at what level a draw should be scored. In sports such as football, a draw is worth 1 point while a win is worth 3. In Chess, a draw is worth a half point versus a full point for a win. Using the half point that is used in Chess, if Player 1 played all of their matches to a draw, they would have no chance of making the top cut, ending up with a rank somewhere in the 40, possibly as low as 49th:

RANK    PLAYER          PTS     SOS     ESOS
49.	Player	1	12	1.55556	1.90741

This should be a sufficient penalty to discourage playing for a draw in a large tournament. However, at a tournament with lower participation, for example an 8 player local Chainbound event with no top cut, Player 1 employing the draw strategy would result in a third or fourth place finish.

RANK    PLAYER          PTS     SOS     ESOS
1.	Player	2	12	1.77778	2.29630
2.	Player	5	10	1.55556	2.22222
3.	Player	1	6	2.44444	2.00000
4.	Player	3	6	2.44444	2.00000
5.	Player	4	6	2.00000	1.92593
6.	Player	6	4	2.44444	1.62963
7.	Player	7	4	1.77778	1.85185
8.	Player	8	0	1.55556	2.07407

Changing the draw award to a quarter point would result in Player 1 finishing in seventh place, which is perhaps too great a penalty for someone who is inexperienced, or plays more slowly by nature.

1.	Player	2	12	1.55556	2.03704
2.	Player	5	9	1.22222	2.00000
3.	Player	3	5	2.11111	1.74074
4.	Player	4	5	1.66667	1.70370
5.	Player	6	4	2.33333	1.40741
6.	Player	7	4	1.55556	1.59259
7.	Player	1	3	2.11111	1.66667
8.	Player	8	0	1.44444	1.85185


Partial point scoring from draws provides a great enough deterrent to discourage playing for a draw and hoping to make a top cut finish.

Final Conclusions

The largest problem with the current KeyForge Going to Time rules is that they create two paths for winning a tournament round: Forge three keys or be in a superior position when time is called. This second condition can be easily exploited by unscrupulous players and should be addressed. While there are other options that could be used to mitigate the exploit, one way that would require minimal changes to the ruleset, require no additional equipment, and disincentivize slow playing as a path to tournament advancement would be to have a hard time limit for each tournament round. If the players cannot complete their match within the allotted time, they are both awarded a half tournament point. This approach offers a minimum amount of disenfranchisement to a superior player, while also not incentivizing a player to play for a draw hoping to advance to the top cut.

Further Research Needed

This experiment focused only on the first stage of tournament play, it does not address what happens when going to time during single elimination rounds. Further research would be needed in that area to determine how a more definitive going to time process could work during those rounds. In a pure Swiss tournament, utilizing draws is a viable alternative to using the written going to time rules.


Blinkingline is a contributor to Archon Arcana and can often be found answering rules questions, judging tournaments, or drinking whiskey. Feel free to reach out to him on the Archon Arcana Discord.