Dominion Strategy Forum

Please login or register.

Login with username, password and session length
Pages: 1 2 [3] 4 5 ... 8  All

Author Topic: [Discussion] DominionStrategy Qualifier for 2012 US National Championships  (Read 25763 times)

0 Members and 1 Guest are viewing this topic.

zxcvbn2

  • Golem
  • ****
  • Offline Offline
  • Posts: 193
  • Respect: +87
    • View Profile
Re: Re: DominionStrategy Qualifying Day Results
« Reply #50 on: June 24, 2012, 10:31:56 pm »
0

What I've learned from the ongoing, cyclical debate between those debating points vs. wins.

Multi-player Tournaments: Never again.
Logged
One Day Cup II Champion: qmech III: Rabid IV: Qvist

metzgerism

  • Baron
  • ****
  • Offline Offline
  • Posts: 56
  • Respect: +15
    • View Profile
Re: Re: DominionStrategy Qualifying Day Results
« Reply #51 on: June 24, 2012, 10:40:18 pm »
0

What I've learned from the ongoing, cyclical debate between those debating points vs. wins.

Multi-player Tournaments: Never again.
As long as people want to hold them, they will continue to exist. Which basically means, as long as there will be over-the-table tournaments, this will continue to be a thorn in the side of some people. Most notably, myself.

---

I understand there's a decent amount of derailment going on here, but since zxcvbn and kirian are watching (and I'm sure WW is too):

For those of you that want to see a point system: what's the reason?
(I can sum up the points given, but I'm already biased and don't trust myself to be objective about what I don't agree with - so please, someone answer it for me, then ask the same question about winner-take-all)
« Last Edit: June 24, 2012, 10:44:39 pm by metzgerism »
Logged

blueblimp

  • Margrave
  • *****
  • Offline Offline
  • Posts: 2803
  • Respect: +1519
    • View Profile
Re: Re: DominionStrategy Qualifying Day Results
« Reply #52 on: June 24, 2012, 11:01:30 pm »
+1

Hey, so here is a suggestion about how to balance points in 3p vs 4p. First off, let's make each game zero sum in points. In other words, the point distribution is:

4p game: 3, -1, -1, -1
3p game: 2, -1, -1

The motivation behind this is so that it doesn't matter too much how many games you play, because the average points among all participants remains zero. OK, now, for 4p, the maximum points you can achieve in 4 games is 12. To make it fair for 3p, the maximum points achievable in the games you play should also be 12, so 3p players should play 6 games, ideally.
Logged

timchen

  • Minion
  • *****
  • Offline Offline
  • Posts: 704
  • Shuffle iT Username: allfail
  • Respect: +233
    • View Profile
Re: Re: DominionStrategy Qualifying Day Results
« Reply #53 on: June 24, 2012, 11:39:37 pm »
+2

Why Point system:

My main reason is that the result reflects more skill than luck. Multiplayer Dominion is a very swingy and luck-dependent game. You can just lose to a newbie with bad draws. 4p game is the most severe, as engines are more unlikely to be built. My experience in multiplayer game is that I am more often powerless to win, but I do have more control not to be last, say in a 3p game. Note that if only 4 wins and 3 wins can proceed this reason is probably not relevant.

A slightly less important reason is that point system has less king-making. Basically in winner-take-all, you cannot predict what opponents will do once they find themselves highly unlikely to win. In addition, since in the current format one only faces the same set of opponents, it is unlikely for anyone to have a chance to play suboptimally in order to make his main competitor even less likely to win.

A even less important reason is that a point based system has no problem dealing with 3 or 4 players or ties.
Logged

metzgerism

  • Baron
  • ****
  • Offline Offline
  • Posts: 56
  • Respect: +15
    • View Profile
Re: Re: DominionStrategy Qualifying Day Results
« Reply #54 on: June 24, 2012, 11:42:32 pm »
0

Hey, so here is a suggestion about how to balance points in 3p vs 4p. First off, let's make each game zero sum in points. In other words, the point distribution is:

4p game: 3, -1, -1, -1
3p game: 2, -1, -1

The motivation behind this is so that it doesn't matter too much how many games you play, because the average points among all participants remains zero. OK, now, for 4p, the maximum points you can achieve in 4 games is 12. To make it fair for 3p, the maximum points achievable in the games you play should also be 12, so 3p players should play 6 games, ideally.
I think this is fine as long as 3p pods play 6 games. Not really a point system, and the negative range for 3p is larger than 4p.  There's still a level of inequity involved in how match-ups pan out, but there's only so much you can mitigate. This does the job well by making 3p games accountable within a similar framework as 2p games.

How do you score ties, though?
Logged

Kirian

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 7092
  • Shuffle iT Username: Kirian
  • An Unbalanced Equation
  • Respect: +9371
    • View Profile
Re: Re: DominionStrategy Qualifying Day Results
« Reply #55 on: June 25, 2012, 12:02:19 am »
0

[snip for brevity]
The only distinction made is between players who won, and players who did not win. According to the actual Dominion rules, there's only one prize handed out: victory (shared or outright). Therefore, winner-take-all seems like the LEAST arbitrary system to employ, in my observation and opinion.

Perhaps, but of course those rules only apply to a single game of Dominion.  By the same rules (including Intrigue rules), a single game of (regulation) Dominion cannot determine the winner of a group of larger than 6 players.  Therefore, the victory rules in the rulebook cannot apply to a tournament situation.  QED.

I should be clear that I don't think it was the non-points system that is arbitrary nearly as much as the only four games and the all of those games with the same players problems, not to mention the seating problem.  I actually think all of those trump the winner-take-all problem.  But I should have been more clear on that, I think, since I harped on the 4P WTA initially and mainly.

For those of you that want to see a point system: what's the reason?
(I can sum up the points given, but I'm already biased and don't trust myself to be objective about what I don't agree with - so please, someone answer it for me, then ask the same question about winner-take-all)

In my opinion, a multi-game winner-take-all scenario encourages poor play, especially from the fourth seat.  There have been discussions about the inherent necessity of taking extra risks from the P2 position in 2P games.  That necessity scales with the number of players.  In a 4P game, P4 must be willing to take risks that are poor play but might still win the game.  Buying two unsupported Maps is, I'm sorry, bad play--but it's good enough, on average, to overcome that huge fourth seat deficit.

It also compounds luck.  Consider a series of four games.  A player who wins one (say, the one where he started) and takes second in three is, in my opinion, likely to be a better player than someone who wins two games--the one he started, and the one where he lucked into T5 Maps as P4--and came in fourth in the other two.  It's reasonable to guess that this player would not have won that game without such great luck.  However, that single lucky win bumps him into serious contention.

Both of these can probably be mitigated by more games.

And I see timchen has already noted these, along with the kingmaker argument.
Logged
Kirian's Law of f.DS jokes:  Any sufficiently unexplained joke is indistinguishable from serious conversation.

rrenaud

  • Administrator
  • *****
  • Offline Offline
  • Posts: 987
  • Uncivilized Barbarian of Statistics
  • Respect: +1177
    • View Profile
    • CouncilRoom
Re: Re: DominionStrategy Qualifying Day Results
« Reply #56 on: June 25, 2012, 12:17:03 am »
0

Did the fixed pods bother other players?  It seemed like a nice win.  You know what's better than waiting in a Dominion lobby?  Playing dominion.  Maybe 4 games of fixed player pods games happens as fast as 2 or 3 variable player pod games.

The big problem with it is that you can get really screwed if you are in a strong pod because the players don't rotate. 
Logged

timchen

  • Minion
  • *****
  • Offline Offline
  • Posts: 704
  • Shuffle iT Username: allfail
  • Respect: +233
    • View Profile
Re: Re: DominionStrategy Qualifying Day Results
« Reply #57 on: June 25, 2012, 12:29:54 am »
0

If abandoning the fixed pods is an option then we can try Swiss! after every round (say, have a time limit of 30 minutes), everyone takes seat according to their current score. The score is the matchpoints I talked about, with the sum of seating as tie breaker. The first player sits at 4th seat on table one, second player sits at 3rd seat on the same table, etc. Can play 4-5 rounds.*

*not really suggesting it, probably induces more luck.
Logged

metzgerism

  • Baron
  • ****
  • Offline Offline
  • Posts: 56
  • Respect: +15
    • View Profile
Re: Re: DominionStrategy Qualifying Day Results
« Reply #58 on: June 25, 2012, 01:01:47 am »
0

Why winner-take-all:

1) It is the only result prescribed in the rules. There is no mention of credit for 2nd or 3rd place in the rulebook for Dominion, just winners (tied or outright). This means that there is no official precedent for having any credit given to players who do not win a game.

2) It is what matters in the latter stages of most tournaments and championships. With a few exceptions, Dominion tournaments are won in a matter of a single game, or a best-of series. To crown a champion, neither of these methods take into account any losers, except for some tiebreakers.

3) It discourages players from colluding with certain opponents to gain a beneficial result for both of them. If you've ever watched a major international soccer tournament, you know that the last two games of the pool rounds are played concurrently. This is meant to mitigate a "common-enough-to-make-a-rule-about-it" situation where the participants of one game will both advance to the next round, regardless of the result of the other game, if they play for a known and exact result. It happened in 1982 and the rule was instituted, although the phenomenon is not preventable in soccer's current regular tournament format (Euro 2012's Group C had this problem scenario). With a winner-take-all system & a first-to-X-wins format, this problem ceases to exist. Without a first-to-X-wins format, winner-take-all still helps prevent this collusion between, for example, a first and second place player.

If I am playing a tournament or league of anything, I want to be playing the actual game prescribed (Reason 1), not a modified or watered-down version of it. I also want to be playing under as-close-to-the-same rules as possible from start to finish (Reason 2), because with different incentives come different styles of play. Finally, I will be playing to win; if there is some loophole in the format that will guarantee my progression in the tournament, I will take it, spirit-of-the-game be damned (Reason 3). I've done it before, and I'd do it again, because it's the organizer's job to make the rules, not mine. My goal is to win.

As an organizer, these are the primary concerns I have, which is why a winner-take-all system (ideally with a first-to-X-wins series) is the only proper way to organize a Dominion tournament, in my opinion.

Now, to respond (please feel free to respond to mine - I'm trying to be formal and equal, so please do the same for me):

Why Point system:

My main reason is that the result reflects more skill than luck. Multiplayer Dominion is a very swingy and luck-dependent game. You can just lose to a newbie with bad draws. 4p game is the most severe, as engines are more unlikely to be built. My experience in multiplayer game is that I am more often powerless to win, but I do have more control not to be last, say in a 3p game. Note that if only 4 wins and 3 wins can proceed this reason is probably not relevant.
I agree that Dominion can be a very swingy and luck-dependent game, and it scales upwards in player count poorly. However, long ago we started having best-of and multi-game series for Dominion, and that's basically become the norm in tournament play. This provides a simple solution for this problem, and retains the game in it's complete form - Dominion is swingy, why prevent that through metagaming? It's just a part of the game, and it still takes a decent amount of skill to see which combo is going to provide your best path to victory.

A slightly less important reason is that point system has less king-making. Basically in winner-take-all, you cannot predict what opponents will do once they find themselves highly unlikely to win. In addition, since in the current format one only faces the same set of opponents, it is unlikely for anyone to have a chance to play suboptimally in order to make his main competitor even less likely to win.
Aside from speaking about the current format, I disagree completely, as the kingmaking is worse with a point system. Assuming everyone plays as optimally as possible, and does not "fall on their sword" so to speak, a winner-take-all system would have everyone not in the lead positioning themselves for the best possible path to victory. This gives the player with the lead the unenviable task of trying to end the game without any help from someone playing poorly.

Having a point system changes this completely. Instead of everyone fighting in their own self-interests, someone with the opportunity to end the game in 2nd place might do so, simply to screw over the two people they are leading. The 1st place player gets a lot more points simply because one of their opponents decided that kingmaking them was preferable to trying to win themselves, and in reality could lead to players without a decent endgame strategy gaining a lot of points simply because their opponents thought it preferable to place 2nd next to them.

This opens up the table for late-round collusion, sub-optimal play, and metagaming. If two players can be guaranteed advancement via an alliance, you have now added politics to the game that DXV tried so hard to make non-political.

A even less important reason is that a point based system has no problem dealing with 3 or 4 players or ties.
I recommended a system almost identical to the old one-off BGGDL system. This would be strictly "winner-take-all," cares about the players that you beat (you winning, them not winning), including shared victory. It scales perfectly and ignores ties.

[snip for brevity]
The only distinction made is between players who won, and players who did not win. According to the actual Dominion rules, there's only one prize handed out: victory (shared or outright). Therefore, winner-take-all seems like the LEAST arbitrary system to employ, in my observation and opinion.

Perhaps, but of course those rules only apply to a single game of Dominion.  By the same rules (including Intrigue rules), a single game of (regulation) Dominion cannot determine the winner of a group of larger than 6 players.  Therefore, the victory rules in the rulebook cannot apply to a tournament situation.  QED.
Agreed, a single game of Dominion cannot determine a winner in a larger group. Furthermore, we generally have no framework for a tournament with >3 players/teams per game. But that doesn't mean that the rulebook is unimportant or should be discarded in a larger tournament setting - we should use it as a framework for organizing. As I state above, once you modify the terms of victory, you might not be playing Dominion anymore.

I should be clear that I don't think it was the non-points system that is arbitrary nearly as much as the only four games and the all of those games with the same players problems, not to mention the seating problem.  I actually think all of those trump the winner-take-all problem.  But I should have been more clear on that, I think, since I harped on the 4P WTA initially and mainly.

For those of you that want to see a point system: what's the reason?
(I can sum up the points given, but I'm already biased and don't trust myself to be objective about what I don't agree with - so please, someone answer it for me, then ask the same question about winner-take-all)

In my opinion, a multi-game winner-take-all scenario encourages poor play, especially from the fourth seat.  There have been discussions about the inherent necessity of taking extra risks from the P2 position in 2P games.  That necessity scales with the number of players.  In a 4P game, P4 must be willing to take risks that are poor play but might still win the game.  Buying two unsupported Maps is, I'm sorry, bad play--but it's good enough, on average, to overcome that huge fourth seat deficit.

It also compounds luck.  Consider a series of four games.  A player who wins one (say, the one where he started) and takes second in three is, in my opinion, likely to be a better player than someone who wins two games--the one he started, and the one where he lucked into T5 Maps as P4--and came in fourth in the other two.  It's reasonable to guess that this player would not have won that game without such great luck.  However, that single lucky win bumps him into serious contention.

Both of these can probably be mitigated by more games.
As I stated above, "more games" is generally the norm. However, I disagree with you completely about a 1-3 player with three 2nd places being better than a 2-2 player without - because the objective of the game is to finish with the most VP, and the 2-2 player did that more often than the 1-3 player, that 2-2 player is clearly better at fulfilling the winning objective of the game. Subjectively believing that the consistent 1-3 player is a better player is just that: subjective. The 2-2 player did what was asked - win - more often, and is therefore more likely to win a game.

Expand this to a 400 game series. Is the 100-300 player any better than the 200-200 player?

Also, is there a better argument against winner-take-all in Dominion, specifically, than "2 Treasure Maps?" I'm not a big fan of the card, but everyone still has the same access to the kingdom, right? No matter what you do, you're going to have to choose a strategy that may not pan out. I see Treasure Maps as an indictment against having one-offs, and not against winner-take-all. In fact, Treasure Maps ARE a part of the game, and if you're just implementing a metagaming system to prevent against the 1/12 chance there's a Treasure Map in the kingdom, you might as well just have a rule saying "No Treasure Maps" instead. Seems a little simpler, but still isn't Dominion.

And I see timchen has already noted these, along with the kingmaker argument.
Which I addressed above. I know for a fact that kingmaker would be worse (I'd exploit it, for one) with a point system that gives primary competition credit for non-winning performances. I think, in regards to the kingmaker issue, you're considering the presence of sub-optimal players in a tournament setting. I'm not.

Sub-optimal players exist in the early rounds, but they get eliminated quickly UNLESS you have a system that encourages them to continue playing sub-optimally. That's fine if the format continues in that way, although I'd say you're not playing Dominion anymore, and instead something quite watered down.

Most of the time, halfway through the tournament, you adopt a bracket system and the point system goes completely out the window. You're playing real Dominion again! But you don't necessarily have the best Dominion players in your bracket. And these sub-optimal players could compound the problem by denying a truly strong player advancement through their poor play. It's a slippery slope that leads to either not playing a Dominion tournament as written in the rules, or not having the best players play each other when the games count the most.

Kirian: I agree about the 4p pods being arbitrary. There's nothing about them that doesn't seem that way. I just wish I hadn't misunderstood your comment!
« Last Edit: June 25, 2012, 01:18:11 am by metzgerism »
Logged

timchen

  • Minion
  • *****
  • Offline Offline
  • Posts: 704
  • Shuffle iT Username: allfail
  • Respect: +233
    • View Profile
Re: Re: DominionStrategy Qualifying Day Results
« Reply #59 on: June 25, 2012, 01:47:35 am »
0

Quote
Aside from speaking about the current format, I disagree completely, as the kingmaking is worse with a point system. Assuming everyone plays as optimally as possible, and does not "fall on their sword" so to speak, a winner-take-all system would have everyone not in the lead positioning themselves for the best possible path to victory. This gives the player with the lead the unenviable task of trying to end the game without any help from someone playing poorly.

Having a point system changes this completely. Instead of everyone fighting in their own self-interests, someone with the opportunity to end the game in 2nd place might do so, simply to screw over the two people they are leading. The 1st place player gets a lot more points simply because one of their opponents decided that kingmaking them was preferable to trying to win themselves, and in reality could lead to players without a decent endgame strategy gaining a lot of points simply because their opponents thought it preferable to place 2nd next to them.

This opens up the table for late-round collusion, sub-optimal play, and metagaming. If two players can be guaranteed advancement via an alliance, you have now added politics to the game that DXV tried so hard to make non-political.
I think what you say here is an overall impression, but does not apply here. The key difference is that you are playing the same opponents in 4 rounds. Given the advancing threshold, it's quite likely only one out of the four players can advance, if it's close, even none. It is thus in your best interest to play as optimal as possible. In your example, imagine how it can happen. If it's the first match, are you willing to take 2nd? Certainly not as long as you still have a shot to win. If you have already won one? I think it is similar. If you've already won two? Maybe. But given you have won the first two it probably does not matter for the rest of the players anyway. If you have already won three than it definitely does not matter.

One can play with numbers, but the conceptual point here is that here a player pretty much has to best all other three players in order to advance. Therefore, getting some credit for being 2nd or not does not matter that much when comparing among players in the same pod. However, those points can be very important in order to compare players from different pods.

Now, I've already said elsewhere, but asking players to play for the win under all circumstances is just unrealistic. For example, if I am behind by more than a Province from the leading player and everyone seems possible to get that last Province. If I can, should I get it? According to your thoughts I should not as I will not win. But this is debatable. The current winning player may have something to say if I didn't take it, as if I am really behind it just looks as if I am keeping that province for the next player. Why should I?  Instead, if we play points by positions, I can comfortably take it as it improves my position.

Yeah, I now see your comment "aside from speaking about the current format", so my point is moot. I agree (especially in Swiss tournaments) it is highly possible. But are we trying to change the pods? I thought the waiting and arranging is just too cumbersome.

Quote
I agree that Dominion can be a very swingy and luck-dependent game, and it scales upwards in player count poorly. However, long ago we started having best-of and multi-game series for Dominion, and that's basically become the norm in tournament play. This provides a simple solution for this problem, and retains the game in it's complete form - Dominion is swingy, why prevent that through metagaming? It's just a part of the game, and it still takes a decent amount of skill to see which combo is going to provide your best path to victory.
Quote
2) It is what matters in the latter stages of most tournaments and championships. With a few exceptions, Dominion tournaments are won in a matter of a single game, or a best-of series. To crown a champion, neither of these methods take into account any losers, except for some tiebreakers.
I imagine Dominion is designed mainly as a casual game. So casually speaking I have no objections here. But if we are talking about a qualification to a tournament, I think the purpose of the qualification is to try to find the best candidate to represent our community, and to have the best chance to win the tournament. Since luck is not something a player can control, we should try to find the players that can win not only based on his luck. (This is in response to quote part 2) If the required skill set is the same for the different format of the qualification tournament (which I do think so, while strategy can be a bit different), we should choose the one which measures least luck, instead of the one used by the higher level/official tournaments.

Quote
1) It is the only result prescribed in the rules. There is no mention of credit for 2nd or 3rd place in the rulebook for Dominion, just winners (tied or outright). This means that there is no official precedent for having any credit given to players who do not win a game.
You should probably ask theory to remove the rule that the point counter should be used first. It's not in the rules. And seriously. I don't think you are  allowed to use one IRL.
Logged

metzgerism

  • Baron
  • ****
  • Offline Offline
  • Posts: 56
  • Respect: +15
    • View Profile
Re: Re: DominionStrategy Qualifying Day Results
« Reply #60 on: June 25, 2012, 02:10:38 am »
0

@Tim

Well, unless there's any other points you want to make I think the derailment is over. I'm speaking generally and you're speaking conditionally, and in these conditions that all makes a bit more sense to me.

I still believe that we must rank players by wins first, and (especially with 4 games played) in no scenario will several 2nd place finishes out-do a single 1st place, but I don't really think you're arguing that. Tiebreakers will matter significantly in this format, anyways.
« Last Edit: June 25, 2012, 02:13:11 am by metzgerism »
Logged

Kirian

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 7092
  • Shuffle iT Username: Kirian
  • An Unbalanced Equation
  • Respect: +9371
    • View Profile
Re: Re: DominionStrategy Qualifying Day Results
« Reply #61 on: June 25, 2012, 02:15:50 am »
0

As I stated above, "more games" is generally the norm. However, I disagree with you completely about a 1-3 player with three 2nd places being better than a 2-2 player without - because the objective of the game is to finish with the most VP, and the 2-2 player did that more often than the 1-3 player, that 2-2 player is clearly better at fulfilling the winning objective of the game. Subjectively believing that the consistent 1-3 player is a better player is just that: subjective. The 2-2 player did what was asked - win - more often, and is therefore more likely to win a game.

Expand this to a 400 game series. Is the 100-300 player any better than the 200-200 player?

Ah, now that is actually an interesting question!  A 400 game series has essentially obliterated the ultra-luck factor, as that ultra-luck is likely to be spread around the table, as opposed to one person getting ultra-lucky in one game.  And this ultra-luck extends to a lot of things:

Being the only 5/2 on a good 5/2 board (of which there are a multitude, pick one).
Being the only 5/2 on a board with no 2s or 5s (ultra-luck hits both ways, of course).
Drawing CEEE with Chapel on T3, and other players get CCCE on T4/5.
Winning the first Prize early (where no other cursers or attacks are available).
And many many others.

I think a 400 game series is plenty to spread that around.  The question is one of "how many games is enough?"  I'll posit that if you can meaningfully segregate good players from poor players in N 2P games, it will require at least 2N games to do the same in 4P.  I don't think anyone would buy an argument that even best-of-three is enough to do this in 2P; best-of-five is probably the minimum.  In 4P, a 10 game series is probably OK, though I'm making a guess here.  Alas, my last Stats course was in... 1999.  So yeah.

Quote
I see Treasure Maps as an indictment against having one-offs, and not against winner-take-all.

Sure.  But I see them as an indictment against even best-of-5 in multiplayer.  Let me further posit:  The more games played in a series, the less meaningful a points system is, and the better winner-take-all becomes.

Kirian: I agree about the 4p pods being arbitrary. There's nothing about them that doesn't seem that way. I just wish I hadn't misunderstood your comment!

No problem, I think I was really unclear.  I think it's really the combination that bothers me:  players each playing against only the same people, only four games, and winner-take-all.  Increase the number of games and shuffle the people around and my argument for a points system really starts to fall apart.  This is where a Swiss-like system starts to make more sense, actually, though I still think you're talking about 10+ games to meaningfully segregate the players.

Alas, we probably could spend an entire other thread discussing this; and I'm only spectating, as I can't go to Nationals or Gencon.  And I'm tired, so I'm off to bed. :)  We'll talk more tomorrow evening I'm sure.
Logged
Kirian's Law of f.DS jokes:  Any sufficiently unexplained joke is indistinguishable from serious conversation.

timchen

  • Minion
  • *****
  • Offline Offline
  • Posts: 704
  • Shuffle iT Username: allfail
  • Respect: +233
    • View Profile
Re: Re: DominionStrategy Qualifying Day Results
« Reply #62 on: June 25, 2012, 02:22:32 am »
0

Quote
I still believe that we must rank players by wins first, and (especially with 4 games played) in no scenario will several 2nd place finishes out-do a single 1st place, but I don't really think you're arguing that. Tiebreakers will matter significantly in this format, anyways.
I don't quite understand where this conclusion is from. Isn't the whole discussion about whether we should progress a player have something like 1-2-2-2 rather than 1-1-4-4?
Logged

metzgerism

  • Baron
  • ****
  • Offline Offline
  • Posts: 56
  • Respect: +15
    • View Profile
Re: Re: DominionStrategy Qualifying Day Results
« Reply #63 on: June 25, 2012, 02:40:59 am »
0

@Kirian

And I think whatever discussion we had is done here too (it's working OMG :D)

Yes, I'd prefer to have players in a LOT of games, as you and tim probably remember from the BGGDL. While that blew up out of control because I didn't control the format properly (and the ranking system came into play, effectively replacing the primary goal of league), if I ever run one again, it'd have a similar nature: encourage playing games, discourage metagame elements, and don't require the administrator to be omni-present. In a tournament format like this things would be a little different.

Ironically, at each player count it might be safe to say "first to 3 wins" for a truly interesting and seemingly fair format, but that honestly has more to do with the succession of each win itself: 1 lucky win, 1 good win, and 1 clincher.

Something like that.
Logged

Personman

  • Conspirator
  • ****
  • Offline Offline
  • Posts: 230
  • Respect: +62
    • View Profile
    • My Friendfeed
Re: Re: DominionStrategy Qualifying Day Results
« Reply #64 on: June 25, 2012, 03:52:31 am »
+1

So I noticed that at least in games 1 and 2 of the CieloAzor, michaeljb, farik, pkbrooklyn pod, the "same starting hands" option was not used, despite being called for in the rules. I guess we probably don't care at this point, but since not everyone in my pod was aware of it either, perhaps it should be called out more forcefully for future days?
Logged
My youtube channel. Isoptropic games with commentary!

Forge!!!

  • Conspirator
  • ****
  • Offline Offline
  • Posts: 248
  • Respect: +128
    • View Profile
Re: Re: DominionStrategy Qualifying Day Results
« Reply #65 on: June 25, 2012, 09:12:13 am »
+1

I'm glad no one here was at the qualifier event in Chicago, it would have never gotten finished.
Logged

WanderingWinder

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 5275
  • ...doesn't really matter to me
  • Respect: +4368
    • View Profile
    • WanderingWinder YouTube Page
Re: Re: DominionStrategy Qualifying Day Results
« Reply #66 on: June 25, 2012, 09:24:44 am »
+1

Did the fixed pods bother other players?  It seemed like a nice win.  You know what's better than waiting in a Dominion lobby?  Playing dominion.  Maybe 4 games of fixed player pods games happens as fast as 2 or 3 variable player pod games.

The big problem with it is that you can get really screwed if you are in a strong pod because the players don't rotate. 
They bug me, yes. i guess to elaborate on the last point, if you have a strong pod, they're ALL screwed, more or less (I mean, I guess not entirely]/i]; I feel like I had a strong pod yesterday, and someone still managed to take three). If you get a really weak pod, then even a moderately good player can crush them terribly.
Strength of schedule is really important, which makes me feel it's comparing apples and oranges.

theory

  • Administrator
  • *****
  • Offline Offline
  • Posts: 3594
  • Respect: +6035
    • View Profile
    • Dominion Strategy
Re: Re: DominionStrategy Qualifying Day Results
« Reply #67 on: June 25, 2012, 09:39:39 am »
0

Fixed pods

After watching what happened yesterday, it's absolutely out of the question to do anything but fixed pods.  It would have tripled the amount of time required.

More to the point, only one person can advance, period.  So if you can't beat your pod on any day out of 4 days, you aren't going to be in contention in the end anyway.

This is a more general point that applies to the below: it is unfortunate, but this tournament isn't really designed to find out who is second-best or third-best.  As long as it results in the correct #1 choice, then it's good enough. 

EDIT: I guess this is not actually entirely true, because if #1 - #4 are all equal in strength and drawn together every single day then perhaps none of them advance.  I think that is unlikely, and seeding is the best we can do about that, assuming that 2p dominion skill is even vaguely a 4p skill proxy.

points vs wins

Here's the best example I can see for a points system: 1-1-4-4 vs 1-2-2-2.  Maybe you think the second player here should advance: but note that under pretty much any system other than 4-3-2-1, the first player is going to advance no matter what.  Does everyone really think 1-2-2-2 is clearly superior to 1-1-4-4? 

I worry that if I implement that, the Day 2 results topic will be "omg kingmaking!".  Yeah, you play screwy when you're P4 and go for TMap x2, but you also play screwy when you realize the other person is gunning for second and will happily end it in a losing position.

Note also that neither player could have possibly qualified yesterday.

3p v 4p

Maybe the best way to do this is to make sure 3p games don't happen, somehow.  It is unfair no matter what.  Either you basically can never advance out of a 3p pod, or you get a great chance of advancing. 
Logged

jayarsea

  • Pearl Diver
  • **
  • Offline Offline
  • Posts: 13
  • Respect: 0
    • View Profile
Re: Re: DominionStrategy Qualifying Day Results
« Reply #68 on: June 25, 2012, 09:44:40 am »
0

Re: 3p v 4p

While not a perfect solution, if you could get someone to fill in for a missing player (drafted out of the Great Hall lounge, if necessary) so that people don't play 3 player games...it'd be at least better than the current situation.

Anyway, I appreciate the effort that went into this tournament and regret that I will be mostly unable to participate in the other qualifiers.
Logged

theory

  • Administrator
  • *****
  • Offline Offline
  • Posts: 3594
  • Respect: +6035
    • View Profile
    • Dominion Strategy
Re: Re: DominionStrategy Qualifying Day Results
« Reply #69 on: June 25, 2012, 09:49:48 am »
0

Preliminary results suggest that based on the rules as written, the advancers are:


knaacku - 3x 1st, 1x 3rd
pirate ship economist - 3x 1st, 1x 3rd
Cielo Azor - 3x 1st, 1x 3rd
Personman - 2x 1st, 1x 2nd, 1x 3rd (8 players defeated)
HiWay2Hello - 2x 1st, 1x 2nd, 1x 3rd (8 players defeated)

The rules call for a playoff between Pman and HW2H ... so you two should play each other (you two can decide how many games, I'd recommend at least a best of 3) and report back.  Feel free to both sign up for Day 2.
Logged

WanderingWinder

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 5275
  • ...doesn't really matter to me
  • Respect: +4368
    • View Profile
    • WanderingWinder YouTube Page
Re: Re: DominionStrategy Qualifying Day Results
« Reply #70 on: June 25, 2012, 09:56:50 am »
0

Why winner-take-all:

1) It is the only result prescribed in the rules. There is no mention of credit for 2nd or 3rd place in the rulebook for Dominion, just winners (tied or outright). This means that there is no official precedent for having any credit given to players who do not win a game.
I don't find this relevant, as the rules are not geared toward a competitive setting, and say nothing of series or tournament play. Obviously, if you're only talking about one game at a time, winning is restricted to one place. If you take tournaments to be knockout, everyone is going to agree winner-take-all?

Quote
2) It is what matters in the latter stages of most tournaments and championships. With a few exceptions, Dominion tournaments are won in a matter of a single game, or a best-of series. To crown a champion, neither of these methods take into account any losers, except for some tiebreakers.
Best-of-series by no means have to include only winning. That is precisely what I'm debating. A single game is a lousy format, so I find holding that up as 'precedent' to be... unconvincing. Actually, I find most precedent arguments unconvincing.

Quote
3) It discourages players from colluding with certain opponents to gain a beneficial result for both of them. If you've ever watched a major international soccer tournament, you know that the last two games of the pool rounds are played concurrently. This is meant to mitigate a "common-enough-to-make-a-rule-about-it" situation where the participants of one game will both advance to the next round, regardless of the result of the other game, if they play for a known and exact result. It happened in 1982 and the rule was instituted, although the phenomenon is not preventable in soccer's current regular tournament format (Euro 2012's Group C had this problem scenario). With a winner-take-all system & a first-to-X-wins format, this problem ceases to exist. Without a first-to-X-wins format, winner-take-all still helps prevent this collusion between, for example, a first and second place player.
I have explained many times why this is backwards, but because you seem to keep throwing it out, I'll explain it again. But it gets its own post.

Quote
If I am playing a tournament or league of anything, I want to be playing the actual game prescribed (Reason 1), not a modified or watered-down version of it.
So you have huge problems playing with identical starting hands, using the point tracker, having random kingdoms rather than letting the players agree on a method, using a 25-card black market deck, etc.? However, mostly I want to go with my response to your point 1 here as well.
Quote
I also want to be playing under as-close-to-the-same rules as possible from start to finish (Reason 2), because with different incentives come different styles of play. Finally, I will be playing to win;
You keep saying this, and it's not a relevant factor - we're all playing to win, too! The point is, we have different notions of what winning is.
Quote
if there is some loophole in the format that will guarantee my progression in the tournament, I will take it, spirit-of-the-game be damned (Reason 3). I've done it before, and I'd do it again, because it's the organizer's job to make the rules, not mine. My goal is to win.

As an organizer, these are the primary concerns I have, which is why a winner-take-all system (ideally with a first-to-X-wins series) is the only proper way to organize a Dominion tournament, in my opinion.


Having a point system changes this completely. Instead of everyone fighting in their own self-interests, someone with the opportunity to end the game in 2nd place might do so, simply to screw over the two people they are leading.
This IS their own self-interest ::)
Quote
The 1st place player gets a lot more points simply because one of their opponents decided that kingmaking them was preferable to trying to win themselves, and in reality could lead to players without a decent endgame strategy gaining a lot of points simply because their opponents thought it preferable to place 2nd next to them.

This opens up the table for late-round collusion, sub-optimal play, and metagaming. If two players can be guaranteed advancement via an alliance, you have now added politics to the game that DXV tried so hard to make non-political.
Guess this goes into my next separate post on the topic anyway.
Quote
I recommended a system almost identical to the old one-off BGGDL system. This would be strictly "winner-take-all," cares about the players that you beat (you winning, them not winning), including shared victory. It scales perfectly and ignores ties.
Bolding mine, and bolded because it's a ridiculous claim.
Quote
Agreed, a single game of Dominion cannot determine a winner in a larger group. Furthermore, we generally have no framework for a tournament with >3 players/teams per game. But that doesn't mean that the rulebook is unimportant or should be discarded in a larger tournament setting - we should use it as a framework for organizing. As I state above, once you modify the terms of victory, you might not be playing Dominion anymore.
This ignores the argument though. Basically, this might not be wrong, but if you want to define this as "it's not dominion", then you can't effectively play tournament dominion. I guess you can have knockouts that everyone hates.

Quote
As I stated above, "more games" is generally the norm. However, I disagree with you completely about a 1-3 player with three 2nd places being better than a 2-2 player without - because the objective of the game is to finish with the most VP,
Again, the whole debate is what we're interpreting the objective of the game to be, so you can't just make this claim. Being first falls under the category of having the most VP, but playing to 'maximize the number of players you have more VP than' also fits the bill. I guess my big problem with your arguments can thusly be summed up: you assume that you're correct, which makes your arguments totally ineffectual to people who disagree with you. And there's little point in arguing with people you already agree with.
Quote
and the 2-2 player did that more often than the 1-3 player, that 2-2 player is clearly better at fulfilling the winning objective of the game. Subjectively believing that the consistent 1-3 player is a better player is just that: subjective. The 2-2 player did what was asked - win - more often, and is therefore more likely to win a game.
Your interpretation is subjective as well. Indeed, if you look at sports analytics at all, you'll know that previous win/loss record, analogous to the 'winner-take-all' thing here, is not the most predictive indicator of future success, and some form of score differential (analogous to the placement here), wipes the floor with it basically every time.

Quote
Expand this to a 400 game series. Is the 100-300 player any better than the 200-200 player?
Depends on what the 300 and 200 are. Is a player who gets 1st 19999 times and second 20001 times REALLY worse than the player who gets 1st 20000 times and 4th 20000 times? Because that is what you are saying.

Quote
Also, is there a better argument against winner-take-all in Dominion, specifically, than "2 Treasure Maps?"
Yes. In fact, you quoted some of them here.
Quote
I'm not a big fan of the card, but everyone still has the same access to the kingdom, right? No matter what you do, you're going to have to choose a strategy that may not pan out. I see Treasure Maps as an indictment against having one-offs, and not against winner-take-all. In fact, Treasure Maps ARE a part of the game, and if you're just implementing a metagaming system to prevent against the 1/12 chance there's a Treasure Map in the kingdom, you might as well just have a rule saying "No Treasure Maps" instead. Seems a little simpler, but still isn't Dominion.
The point isn't treasure map. Treasure map isn't even the swingiest card. But it's an exemplar of a card that, straight-up, without help, leads to a strategy requiring little to no skill being able to WIN most often, without actually being able to do well very consistently at all. It's a card that, without support (as an aside, with support, it's a totally fine card - takes a lot of skill to use well), you can have a situation where the game is very directly decided by your Random Number Generator. People don't like that. Now, it's always true to some extent, but TM gives you very little control. But the problem isn't "Treasure Map" per se; TM just represents mindless, luck strategies, where taking risks that on average leave you behind are rewarded.
Quote
Sub-optimal players exist in the early rounds, but they get eliminated quickly UNLESS you have a system that encourages them to continue playing sub-optimally.
Like a winner-take-all format!
Quote

Quote
Most of the time, halfway through the tournament, you adopt a bracket system and the point system goes completely out the window. You're playing real Dominion again!
Okay, you advocate brackets. Nobody wants to play that, so while this is logically consistent, it's not something that should be pursued. It's a game, yes? So we should do what people like, what people want to do. You know, you want it one way, and that's valid. But the point is that if everyone else wants to play Dominion+, then they should play Dominion+, you know?
Quote
But you don't necessarily have the best Dominion players in your bracket. And these sub-optimal players could compound the problem by denying a truly strong player advancement through their poor play.
Like in a winner-take-all format!

WanderingWinder

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 5275
  • ...doesn't really matter to me
  • Respect: +4368
    • View Profile
    • WanderingWinder YouTube Page
Re: Re: DominionStrategy Qualifying Day Results
« Reply #71 on: June 25, 2012, 09:58:27 am »
0

I'm glad no one here was at the qualifier event in Chicago, it would have never gotten finished.
The actual playing went quite quickly. The whole thing was done in about an hour. And the people posting here are largely the same ones who played.

zxcvbn2

  • Golem
  • ****
  • Offline Offline
  • Posts: 193
  • Respect: +87
    • View Profile
Re: Re: DominionStrategy Qualifying Day Results
« Reply #72 on: June 25, 2012, 10:00:27 am »
+1

Quote
I worry that if I implement that, the Day 2 results topic will be "omg kingmaking!".

To be fair, this happened yesterday already in our group. pops had no chance of winning game one (or even taking anything but 4th) and bought the last province before me and methods had a chance at it. We were both down by less than a province, I think I was down one and methods was down three. He bought it because he felt he had the better shot at advancing if the lowest ranked player won. [Log: http://councilroom.com/game?game_id=game-20120624-113610-823a76de.html]

My only point to this is that you can't eliminate kingmaking in any format. This shouldn't really be a reason for choosing one.

Quote
but you also play screwy when you realize the other person is gunning for second and will happily end it in a losing position.

Can someone explain to me how you can gun for second before the endgame in Dominion? This might be relevant in other games, but I don't see how this can happen in Dominion.

.

With all that said, I honestly would prefer we keep the same tournament format throughout. A question about the top 16 however. Will that be single elimination? All it says on the front page is that they will "duke it out." Does this mean 4p single elimination?
« Last Edit: June 25, 2012, 11:07:49 am by zxcvbn2 »
Logged
One Day Cup II Champion: qmech III: Rabid IV: Qvist

WanderingWinder

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 5275
  • ...doesn't really matter to me
  • Respect: +4368
    • View Profile
    • WanderingWinder YouTube Page
Re: Re: DominionStrategy Qualifying Day Results
« Reply #73 on: June 25, 2012, 10:08:20 am »
0

Fixed pods

After watching what happened yesterday, it's absolutely out of the question to do anything but fixed pods.  It would have tripled the amount of time required.
I don't see evidence for this, and I suspect it's a big exaggeration, but whatever. It will take longer of course, and I understand that given our constraints, logistical concerns must trump other concerns.

Quote
More to the point, only one person can advance, period.  So if you can't beat your pod on any day out of 4 days, you aren't going to be in contention in the end anyway.
Are you trying to impose this as an additional constraint? Because it seems EMINENTLY plausible to me that a pod has two people who go 1-1-2-2 or maybe a 1-1-2-3 or something, and that gets them both through some day, if 1-per-pod is not an actual rule.

Quote
This is a more general point that applies to the below: it is unfortunate, but this tournament isn't really designed to find out who is second-best or third-best.  As long as it results in the correct #1 choice, then it's good enough. 

EDIT: I guess this is not actually entirely true, because if #1 - #4 are all equal in strength and drawn together every single day then perhaps none of them advance.  I think that is unlikely, and seeding is the best we can do about that, assuming that 2p dominion skill is even vaguely a 4p skill proxy.

points vs wins

Here's the best example I can see for a points system: 1-1-4-4 vs 1-2-2-2.  Maybe you think the second player here should advance: but note that under pretty much any system other than 4-3-2-1, the first player is going to advance no matter what.  Does everyone really think 1-2-2-2 is clearly superior to 1-1-4-4? 
Except that in almost EVERY point system, 1-2-2-2 advances over 1-1-4-4. 4-3-2-1, 5-3-2-1, 6-3-1-0, 4-2-1-0, 8-3-1-0... in short, every system that's been discussed sees 1-2-2-2 being better. So... I'm calling that out - you have that backwards, and apparently haven't done the calculation there? Or one of us is calculating wrong. But I'm fairly sure I've got the calculations down here.
I don't think 1-2-2-2 is clearly superior to 1-1-4-4, but that's just because nothing is clear after 4 games :P. In general though, I do think that set 1 of the results does indicate superior performance, yes.

Quote
I worry that if I implement that, the Day 2 results topic will be "omg kingmaking!".  Yeah, you play screwy when you're P4 and go for TMap x2, but you also play screwy when you realize the other person is gunning for second and will happily end it in a losing position.
This will be addressed in a soon-to-come-post.

Quote
Note also that neither player could have possibly qualified yesterday.

3p v 4p

Maybe the best way to do this is to make sure 3p games don't happen, somehow.  It is unfair no matter what.  Either you basically can never advance out of a 3p pod, or you get a great chance of advancing. 
I like this, if you can figure out the 'somehow'.

theory

  • Administrator
  • *****
  • Offline Offline
  • Posts: 3594
  • Respect: +6035
    • View Profile
    • Dominion Strategy
Re: Re: DominionStrategy Qualifying Day Results
« Reply #74 on: June 25, 2012, 10:11:01 am »
+1

Apologies on the points system.  I was thinking of another example.

I'm happy to try out points for tomorrow, and see how that affects things.

Quote
Are you trying to impose this as an additional constraint? Because it seems EMINENTLY plausible to me that a pod has two people who go 1-1-2-2 or maybe a 1-1-2-3 or something, and that gets them both through some day, if 1-per-pod is not an actual rule.
I meant, only one person will ever advance from this tournament.  Obviously more than 1 can advance from the same pod.
Logged
Pages: 1 2 [3] 4 5 ... 8  All
 

Page created in 0.158 seconds with 20 queries.