June Tournament Format Poll

[Cognis]

No matter what I think it should be at least best of 3 to minimise the luck factor. You get mana screwed and the AI gets a good start and you've lost before it even started.

[aww1979]

The format votes are fairly evenly split. I think I'm going to make an executive decision and go with the elimination style tournament for June. If it turns out to be a disaster, we'll do something else next.

[foolosopher]

Hi guys, after months of ghosting the forums, I am putting my first post here, so a big thanks first to all of you that contributed and kept the game alive and evolving. Now my quick questions.
1)When and to who are we going to send our decks to participate in the tourne?
2) What happened to the 2000 card limit, is it deemed unsolvable, because I don't see it in "Unsolved Mysteries" and I was planning to look into it?

[aww1979]

1) There's a tournament for June ongoing right now; it's in the semifinals right now, and there will be a new one start probably in a week and a half or two weeks or something like that, depending on when the current one is done and how long it takes me to set it up.
2) It's mostly unsolvable. There is a 'my2000.exe' or something like that, which makes it so that you can choose which 2000 cards to use (out of about 2300 coded) but you still cannot use more than 2000 at once in your decks. There is theoretically a way to solve it, but it requires a) knowing coding very well and b) going through the entire program in assembler and tediously finding every single pointer to the 2000 card array and changing it to point to a bigger array. Snacko knows how to do it, but said that it would take a ridiculously long time and that it isn't worth it for him; I asked my brother, a professional coder, and he said basically the same thing Snacko did in his explanation (back like a year or more ago)

[jatill]

Now that there is a July tournament confirmed, I'll post my thoughts on the June tournament scoring scheme.

Pros:
-It was definitely "fair", in the sense that it gave each deck the same number of chances to score points
-The symmetry of playing each deck from both halves felt good
-Testing was fast

Cons:
-Each deck got to compete only against a limited number of other decks (from 1 to 4)
-Brackets don't scale well when there are a number of players just over 16
-One bad matchup means your tournament is over

To me, the first con outweights everything else. It made the tournament unfun for me in 2 ways. First, there was no variety in my matches. 2 Progenitus mirrors and done. I would much rather have seen how my decks fares against the field instead of the mirror twice. There was literally no opportunity to play against another deck in the tournament. Second, it made the tournament reports boring to read (so I skipped them this time). Reading how the decks play out against each other is fun for me, but reading the same matchup report 13 times is not. So for the sake of variety alone, I hope the scoring format changes to something different for the next round.

[Cognis]

I think it would be better if we divided the decks in 4 groups by 4 and then the top 2 progressed into the quarter finals. It would mean more matches to be played but that wouldn't be such a big problem. Even if there are 17 or some other odd number of decks, one group would have say 5 decks and still the two best would proceed. The scoring is definitely ok so I wouldn't change it.

[aww1979]

Cognis's idea is pretty good. (in fact, it is the world cup idea, just less people voted for it) It would also remove all of jatill's concerns, though having over 16 players would still be a bit tricky to handle, but like cognis said we could probably just have some groups take an extra deck or something.

I'm also thinking that perhaps the points for AI victories are too high. If your deck loses to the AI, whether 1-2 or 0-2, not only have you been set back the one or two points you could have earned if you'd won, but your opponent gets enough points that the only way to catch them is if their deck loses to yours when yours is the AI. It ends up coming down to basically the number of times the AI won. The general scoring idea seems good, but I'll most likely drop the AI point rewards from 3 (for 1-2) and 5 (for 0-2) to something lower.

[jatill]

I like the AI score part, since that it really the make or break factor for most of these decks. Otherwise I'll just build stupid combo decks that the human should never lose with I like the world cup scoring method as a trial.

[juzamjedi]

The AI should definitely get extra points for winning. It makes a ton of bad play choices (e.g. it doesn't counter Natural Order with Force of Will, etc.) so if it manages a win then give him extra credit. You could make minor tweaks if you like and look backwards at results to see what the impact would have been, but overall I think giving AI extra credit is a good thing.

[Cognis]

How about that the AI gets 2 points for each win and the player gets 1 and the winner of the match gets an extra point.

Player wins 2-0 - 3 points, 2 for 2 wins, 1 game
Player wins 2-1 - 3 points player, 2 for 2 wins, 1 for game, 2 points AI for 1 win
AI wins 1-2 - 1 point player for the win, 5 points player for 2 wins and game
AI wins 0-2 - 5 points AI for win and game

I think the points players would get would be better balanced this way, but I really think that 1-2 result should stay 3 points to the AI and no points for the player, while 0-2 result should stay 5 points to the AI and no points for the player.

[aww1979]

Well, the tweak I was thinking was that for when you lose to the AI (1-2 or 0-2) the reward would be dropped maybe by one point for each of those, so the AI deck would get 2 if it won 1-2 and 4 if it won 0-2. (rather than 3 and 5 now)

The way it is right now, and what I forgot to consider when making the system, is that if your deck wins as the AI, not only are you gaining extra points (3 or 5) but you are *denying* your opponent 1 or 2 points because they didn't win. I was just thinking of doing it the same way, only make the 3 and 5 into a 2 and a 4.

[juzamjedi]

Aww1979: how would the proposed scoring system change the results of matches? What value(s) would make the change closer to how you think it should end up?

I'm OK with tweaking the scoring as long as AI wins make some kind of significant impact. The main thing that keeps things interesting is giving extra points to AI wins (because the AI plays like my 9 year old )

[discoransom]

I think the 1-2 record should be 2 points for the AI, and the 0-2 record should remain 5 points. It splits some of the difference in opinion, and still provides a significant AI reward.

[aww1979]

Looking at the scores for the tournament, the 2-4 system (rather than 3-5) would not have changed who won in any cases. Most games were won by more than 7 points, the maximum you can earn from one player's test. Only five matches were closer than 7 points, and even those would not have been changed by my proposed rule (though some would have been a few points closer) so I am now proposing that we leave the scoring system as is.

One change I did think of, that will make my life a bit easier. If in the future there is a tie, instead of playing a sudden death round, I will instead compare scores for previous rounds and use the differential to determine who advances. If that is also a tie, I will still play the tiebreaker. Sudden death will also still be necessary if the tie is in round 1.

As an example, suppose I tie with Jon Finkel in round 2. However, in round 1, I won my match by 5 points, 30-25 whereas in round 1, Finkel won his round by 20 points, 40-20. Finkel would advance to round 3 and I would be out.

July will also follow Cognis/Brandon's suggestion of doing the world cup style round 1, where there will be groups (of 4 or whatever convenient number comes up) and the top two in each group (by points earned) advance to round 2.

[thedrigo]

July will also follow Cognis/Brandon's suggestion of doing the world cup style round 1, where there will be groups (of 4 or whatever convenient number comes up) and the top two in each group (by points earned) advance to round 2.

I like this better, for sure.