Collectible Card Game Headquarters

This thread is for discussion of the format for the June tournament. Options 1 and 2 are the way we ran things in April and May, except that we'd insist on keeping to best of 1, or best of 3, to keep things more even in testing.

Explanation of elimination style tournament:

Finkel, Budde, Rosewater, and Justice each submit a deck. The tournament coordinator (me) randomly determines who they will face in the first round, and finds that in round 1:

Finkel plays Budde
Rosewater plays Justice

For the Finkel-Budde matchup, everyone would play a best of 3 duel with a human Finkel and AI Budde, and a second duel with an AI Finkel and human Budde. (so that both decks are played by both sides) The same concept applies to the Rosewater-Justice matchup, and any other matchups (for more players than this example). There would likely be a two day window for everyone to play a particular matchup, i.e. the Finkel-Budde matchup would be June 1-3 and the Rosewater-Justice matchup would be June 4-6. Everyone reports the result of their best of 3 duel, both with AI finkel and AI budde. Whichever deck accumulates the most wins (for the human) advances to the next round.

So, for the Finkel-Budde matchup, the four players each play with Finkel's deck in a best of 3 duel, giving Budde's deck to the AI. The scores are:
Finkel 2-0
Budde 2-1
Rosewater 1-2
Justice 2-1
total: 3 humans win (out of 4)

They also play with Budde's deck, giving Finkel's to the AI and get:
Finkel 2-0
Budde 2-0
Rosewater 2-1
Justice 2-1
total: 4 humans win (out of 4)

(if there had been a tie, we would then look at the number of game losses. If one deck had all testers go 2-0 and the other deck had all testers go 2-1, the one with 2-0 is the winner. If that's still a tie... I'll think of something later :p)

Budde's deck wins 4-3 against Finkels, and Finkel's deck is eliminated from the tournament, while Budde's deck advances. (though Finkel may continue to test decks)

The same procedure is then followed for the Rosewater-Justice match. Suppose that Rosewater wins. Then, scheduled for June 7-9, would be the Budde-Rosewater match, and the winner of that would be the tournament winner. Then, all players would test Budde and Rosewater in exactly the same manner.


This could be modified somewhat, such as best of 5 games, or whatever, but the basic idea is there. Other ideas, or modifications of this or the gauntlet format done in the last two months, are welcome, and will be added to the poll.

If a number of entries other than 4,8,16,32, etc are received, I will either add my own decks to fill in the spaces and/or give some players a 'bye' which means that they will not have to play in the first round, and automatically start in the second round. Byes would be either determined randomly, or maybe based on performance in the previous tournament.

You could make all of the pairings go into "1 bracket" instead of having "divisions". This reduces the need for multiples of 4.

Also want to say I am OK with any format Vintage, EDH or otherwise.

We could have Tournament like football World Cup.

Group A
Deck 1
Deck 2
Deck 3
Deck 4

Group B
Deck 5
Deck 6
Deck 7
Deck 8

and so on...

Best two of each group advance to playoffs. Group A winner plays against Group B Second and Vice Versa...and eventually semi-finals and final.

Points in group phase could be like this: 2-0 win three points. 2-1 win two points for winner and one point to loser. IF two decks have same amount of points, the winner of their mutual match would be higher.

In one group would be 6 games. If 16 decks participate then it would take 24 games in four groups. 8 decks in to playoffs so 4 more games plus 2 in semifinal and final. Total 31 games.

Points could go like this: 1st=12 points, 2=8 pts, 3=6 pts, 4=4 pts. Every top 8 loser would get 2 points. So if deck goes through to playoffs it's guaranteed 2 points at least.

Human vs. AI. So didn't think this side when I started to write, but we could propably work something out.

Something else: I think that 10 round gauntlet isn't the way to go anymore and we need some type of tournament or league mode in future.

my suggestion:

8 decks total (more would consume way too much time)

playing a best of 3 against each other.means EVERY deck got the same chance for a matchup.Best of 3 should keep the luck factor even lower.

so you got 1 deck , and play a best of 3 seven times against all the other decks one by one (no mirror).
would equal a minimum of 14 games/deck and a minimum of 112 games total.

optional:
playing 1 deck per day and create a report on the forum.this way we could do like 3 tournaments/month.
also a fixed time limit of 1 hour per best of 3 might be a good idea.(game is a tie then)

the deck with the best overall (ai+human) win/loss ratio wins.
this should eliminate the "unfair" advantage of decks getting more chances at free points while not getting any penalties for it and also should be superior to a highly luck based best of 1 scenario.

Two new options added to poll. You can change your vote if desired.

even though it stinks a bit for those of us with less time on their hands I support the 8 decks maximum (first come first served, though 1 deck/day or three tournaments/month is clearly insanity) and best-of-three versus every deck gauntlet. which only makes sense if all of those who enter decks actually do the testing.

seeing your own deck perform is a big eye-opener with the AI (first time around I did not test my deck in the hands of the AI) and something that will help you next time around. and why not "reward" the people who put time into these challenges?

so I vote "other", the Salbei option for sane people to be specific. time limit is a good idea but I would lower it to 45 minutes or even less.

My preference really depends on the number of players.
I am not in favor of limiting the number of participants. We should be able to adapt the rules in order to avoid tedious plays (and with a high number of players, the luck factor automatically decreases).

Edit : we could also add a special prize. Each player votes for the deck he enjoyed the most in the tournament (voting for his own deck is forbiden), and the participant with the most popular deck can also chose a card to be coded.
What is your opinion ?

@yggdrasil: I was actually already thinking of doing something like that, so I have no problem at all with the idea

I'm for the option of everybody playing everybody with an elimination round. If there are few decks we might even make it best of 5 to minimize the luck factor but best of 3 is more doable if there are plenty of players.

Let's say that is the first round. After we all test the decks that way we just add the points 1 points for a won match by the player and 2 or possibly 3 points for the AI if it wins. After that we tally the scores and the first 8 or 4 or 16 progress to the elimination best of 5/7/9... where the 1st deck would play the 8th and so on.

Then we would have to award points when the final standings are done (etc 1st place 10 points, 2nd 8 points - F1 style or something similar We again tally the scores and get the overall winner.

This might sound a little complicated but even if we play best of 1 in the qualifying stage the best decks will get to the top 8 since the luck factor is minimal and even the loss from the AI would allow you to get points. We could even allow something like substitutions or a sideboard players could exchange before the elimination part, but that might complicate things, for now I think we should just pick the way we will score.

At this point, I'm leaning in favour of either the elimination-style tournament or else salbei's everyone plays everyone idea. I like salbei's idea, but it would take longer. For instance, according to my statistics, the average player played 52 matches, but if in that same tournament, we had each played everyone, we would have played (14-1)*(14-1)=169 matches, just over triple the amount. (the -1 excludes the mirror match, which is pointless) If we played best of 1 instead of best of 3, it would work out to roughly the same number of actual games, though. With my idea, with 14 players, the elimination style tournament would have had 7+3+2+1=13 matches, which is very fast, but means many matchups never occur.

Then it would be nice to have a winner/loser bracket, with a finale between the 1st deck of the winner and the loser, although there will be more match to be played

Looking at the results of the last tournament, I see that even the unluckiest deck was still played by the AI 37 times. I think that's a significant enough number to get a win % from. How about we just use overall AI win% as the grading factor? I'd recommend 10 points for a 10-round gauntlet again, and then up to 30 points for AI win% (100% wins = 30 points, 50% = 15 points, etc).

(If there were 7 testers, you would have 7 different gauntlet scores, but just one overall AI win% score, which would get multiplied by 7)

I think that would work no matter how many players we would have.

Just another idea, how about we throw out mirror matches? If you get a mirror match just ignore it so players couldn't get points for loosing to their own deck. In a real tournament you can't play against yourself so I don't see much point you would play here. In some of my gauntlets I played 2-3 mirror matches which is a pain in the ass since you want to test the deck against others.

I voted for Elimination style. In this format, the required effort for playing out the tournament is relatively low (even with a high number of participats), but still, but still, the luck factor is relatively low (the only way you could get unlucky would be to be paired against an unfavorable matchup with an otherwise strong deck). For similar reasons, Brandon's idea would be my secod preference.