Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Rabid]

Please, do not use the same number thoughout the tournament. That would mean that someone with a really low number would ALWAYS go first. It should be random every match.

Do you need to re randomise each round?
I'm not sure, it feels wrong not to, but this could be a statistics thing that feels unfair but isn't?

[greatexpectations]

Bah. I hate the PCE and would much rather prefer it be banned if one player in a match requests so.

sure, but as we have said we can't enforce anything. ideally, the approach to this situation would be similar to what you see from other competitions with strong DIY or community based roots: the 'Spirit of the Game' approach to ultimate and the 'fair play' concept in soccer both come to mind.

the spirit of the game wiki discussion summarizes it succinctly:
"Highly competitive play is encouraged, but never at the expense of the bond of mutual respect between players, adherence to the agreed upon rules of the game, or the basic joy of play."

[Lekkit]

I didn't really want to spark a discussion, just get my opinion said. And I don't agree that you have to use it or be at a disadvantage. You have to know how to use it or be at a possible disadvantage. I won't install it for this tournament. But I'm pretty sure I will type !details at least once a turn in the games where my opponent wants to play with it on.

[Lekkit]

Please, do not use the same number thoughout the tournament. That would mean that someone with a really low number would ALWAYS go first. It should be random every match.

Do you need to re randomise each round?
I'm not sure, it feels wrong not to, but this could be a statistics thing that feels unfair but isn't?

I'm not super into statistics, but I think it would be something like randomising starting hands for each player at the start of the tournament and then always use those hands. Someone will get a 5/2 and keep it thoughout the entire tournament. I think it would be really unfair. Win the lottery once, always win the lottery.

[greatexpectations]

I didn't really want to spark a discussion, just get my opinion said.

hey discussion is fine. we just want to keep things friendly and civil. as the mods for this, we get stuck in a rough spot. id assume this is why theory wanted assistance for this tournament, as he was likely frustrated by the drama surrounding the last tournament he organized.

we will do all we can to try to minimize luck and keep things fair, but unfortunately there is a mix of human element in all this too. reducing potential luck factors can mean little when the tournament is almost entirely anonymous. at some point we have to do what we feel is best and then just leave it at that.

[jonts26]

Please, do not use the same number thoughout the tournament. That would mean that someone with a really low number would ALWAYS go first. It should be random every match.

Do you need to re randomise each round?
I'm not sure, it feels wrong not to, but this could be a statistics thing that feels unfair but isn't?

I'm not super into statistics, but I think it would be something like randomising starting hands for each player at the start of the tournament and then always use those hands. Someone will get a 5/2 and keep it thoughout the entire tournament. I think it would be really unfair. Win the lottery once, always win the lottery.

Imagine for simplicity sake a tournament in which every player played 8 rounds. If we randomize each round, you have a 1 in 256 chance of going first all 8 times (2^8) or second all 8 times. With pre randomizing you have a 1 in 256 chance of getting the 1 seed. So the odds are the same. It's just we pushed them all up front in one case. The math gets trickier with elimination and the other seeding, but it should still work out, either way that your expected number of series where you go first overall is 50%. The only thing that changes is when you know how lucky you got.

Anyway I guess it doesnt matter. GE wants to just make this as simple as possible so I dont have to explain the math over and over. So we're going to just randomize each round since it is more intuitive and very little extra work.

[Fabian]

The expectation might be the same, but the variance certainly isn't. Definitely a new number each round please.

[DStu]

The expectation might be the same, but the variance certainly isn't. Definitely a new number each round please.

Disclaimer: I'm not taking part in the tournament.

You can certainly construct it (quite easily) the way that it is equivalent (up to the fact that you know beforehand).  The way it was proposed it's probably not, as if you go first somehow depends on who's advancing.
But you could e.g. just do all the coin tosses beforehand, and than construct the bracket so that it is equivalent. 4 player example:
Take player A vs B and C vs D, and 3 cointosses.  First take the coin for the final, if winner of A vs. B goes first, set first bit of A's and B's number to 0, otherwise 1. C and D get the opposit.  Now take A and B, toss coin, if A goes first, second bit of A's number is 0, otherwise 1. B the opposite. C and D analogue.  Now if you just reveal the last remaining bit each round, from the perspective of the players it looks like you would toss the coin for each round.  But nevertheless you have done all coin tosses beforehand.

[Varsinor]

The expectation might be the same, but the variance certainly isn't. Definitely a new number each round please.

Not even the expectation value is the same if I am not mistaken. When the coin is tossed every time, your expectation value to go first is 1/(2^number of matches) or 1/256 for 8 matches.
Under the proposed system, this is guaranteed for one out of 256, the player with number 1. But the next, say, 9 players are also very likely to go first on every game because it is unlikely that they will face a player with a lower number directly.

With this system, you'll end up having plenty of people who have gone first on every match and plenty of people who have gone last and way too few people who have gone first every second match like they should have.

I don't really like this. I can also understand the organizers' desire to keep the amount of work low, of course! But in case it is not that much more work, I would really prefer tossing the coin for every match...

[jonts26]

The expectation might be the same, but the variance certainly isn't. Definitely a new number each round please.

Not even the expectation value is the same if I am not mistaken. When the coin is tossed every time, your expectation value to go first is 1/(2^number of matches) or 1/256 for 8 matches.
Under the proposed system, this is guaranteed for one out of 256, the player with number 1. But the next, say, 9 players are also very likely to go first on every game because it is unlikely that they will face a player with a lower number directly.

With this system, you'll end up having plenty of people who have gone first on every match and plenty of people who have gone last and way too few people who have gone first every second match like they should have.

I don't really like this. I can also understand the organizers' desire to keep the amount of work low, of course! But in case it is not that much more work, I would really prefer tossing the coin for every match...

1. I've already stated that we are switching to a per round randomization.
2. What you are describing is not a change in the expectation, but a higher variance, as Fabian pointed out. Expectation is still 50% going first, but it's skewed so you get more extremes like going first/second a higher proportion of the time. But yes, this is reason enough to use the per round randomization.

[Varsinor]

I'll include these instructions again when we announce brackets. You will have to play your first game in the tournament lobby, selecting who goes first. Then, both players can either manually set up each game, or head over to the Secret Chamber lobby where loser goes first will be preserved.

I am not quite sure if you mean that after the first game, players may rotate in being player 1 (independently of who won) or if after the first game, the looser always goes first which would seem to indicate that after draws, you can't play in the Tournament lobby but must change to the Secret Chamber.

I would prefer the former.

[Varsinor]

1. I've already stated that we are switching to a per round randomization.

Ah, sorry, I missed that. Great!

2. What you are describing is not a change in the expectation, but a higher variance, as Fabian pointed out.

Indeed, absolutely! My mistake, I stand corrected! (I daresay that usually doesn't happen on statistics questions as simple as that, I'm kind of embarrased now... ) I actually kind of implied it myself in the last-but-one paragraph of my posting... I guess I wanted to fire off that posting too quickly!

[jonts26]

I'll include these instructions again when we announce brackets. You will have to play your first game in the tournament lobby, selecting who goes first. Then, both players can either manually set up each game, or head over to the Secret Chamber lobby where loser goes first will be preserved.

I am not quite sure if you mean that after the first game, players may rotate in being player 1 (independently of who won) or if after the first game, the looser always goes first which would seem to indicate that after draws, you can't play in the Tournament lobby but must change to the Secret Chamber.

I would prefer the former.

We are playing by 'standard isotropic rules.' First player in the series is randomized, after which loser goes first. In the event of a draw, turn order is preserved from the previous game.

The easiest way to ensure this is just having two players play in the secret chamber lobby. But since I am randomizing the order for first turn overall, we need to set that manually, which must be done in the tourney lobby. After that, feel free to go back to the SC lobby to finish the series. If you want to stay int he tourney lobby, feel free, but you will have to manually set who goes first each game. Which is still loser first per the tournament rules.

[Watno]

If the expectation value of games you go first in wouldnt be half the number of rounds, something would be kinda strange.

[Varsinor]

We are playing by 'standard isotropic rules.' First player in the series is randomized, after which loser goes first. In the event of a draw, turn order is preserved from the previous game.

Are you sure these are the isotropic rules regarding draws? Has Dougz posted that somewhere?

His two postings here (*click*) sound like in case of another 2-player game after a draw, the starting player is randomized and not determinded according to the previous game.
Both players fall into the same class of players after the draw. I would guess the class is "winners", because Isotropic (and the original rules) say both players win on a draw. But even if it counts as a "conventional draw" and therefore as the class "non-winners", both players are still in the same class.

[jonts26]

We are playing by 'standard isotropic rules.' First player in the series is randomized, after which loser goes first. In the event of a draw, turn order is preserved from the previous game.

Are you sure these are the isotropic rules regarding draws? Has Dougz posted that somewhere?

His two postings here (*click*) sound like in case of another 2-player game after a draw, the starting player is randomized and not determinded according to the previous game.
Both players fall into the same class of players after the draw. I would guess the class is "winners", because Isotropic (and the original rules) say both players win on a draw. But even if it counts as a "conventional draw" and therefore as the class "non-winners", both players are still in the same class.

OK, I'm not positive what the iso rule is. If anyone knows speak up. REGARDLESS, we are using the iso implementation, which is either order is preserved or order is randomized. Either way, it should have minimal impact on results.

And to pre-empt you, after a draw you have to go to the secret chamber lobby.

[Mic Qsenoch]

I don't actually care about this issue at all, but just so everyone knows, I tested it and in case of a tie the person to start the next game is determined randomly.

[theory]

1. Point counters are permitted.  If you don't want to play with them you can ask your opponent not to use them and it is up to them whether to listen to you.

2. We'll edit the rules.  In the first game of the series, the person "on top" in the Challonge bracket is designated as going first.  Afterwards, loser goes first.  In event of a tie game, immediately replay, and use Isotropic rules where first player is randomly determined.

[blueblimp]

Please, do not use the same number thoughout the tournament. That would mean that someone with a really low number would ALWAYS go first. It should be random every match.

For any way that first players are chosen and for any tournament results, it's always true that there is some player that starts first in every series he/she plays. Try it out.

(Proof idea: Suppose each player started second in some series. Then every player who lost in the first round must have started second, since otherwise they would have started first in their only series. So, every player who won in the first round must have started first. Similar reasoning works for the second round and so on, but then the tournament winner could never have started second.)

[jonts26]

Please, do not use the same number thoughout the tournament. That would mean that someone with a really low number would ALWAYS go first. It should be random every match.

For any way that first players are chosen and for any tournament results, it's always true that there is some player that starts first in every series he/she plays. Try it out.

(Proof idea: Suppose each player started second in some series. Then every player who lost in the first round must have started second, since otherwise they would have started first in their only series. So, every player who won in the first round must have started first. Similar reasoning works for the second round and so on, but then the tournament winner could never have started second.)

False.

Consider the case where in the first round, every player who started first lost.

[blueblimp]

Please, do not use the same number thoughout the tournament. That would mean that someone with a really low number would ALWAYS go first. It should be random every match.

For any way that first players are chosen and for any tournament results, it's always true that there is some player that starts first in every series he/she plays. Try it out.

(Proof idea: Suppose each player started second in some series. Then every player who lost in the first round must have started second, since otherwise they would have started first in their only series. So, every player who won in the first round must have started first. Similar reasoning works for the second round and so on, but then the tournament winner could never have started second.)

False.

Consider the case where in the first round, every player who started first lost.

If you start first in the first round and lose, then you started first in every series you played.

[jonts26]

Please, do not use the same number thoughout the tournament. That would mean that someone with a really low number would ALWAYS go first. It should be random every match.

For any way that first players are chosen and for any tournament results, it's always true that there is some player that starts first in every series he/she plays. Try it out.

(Proof idea: Suppose each player started second in some series. Then every player who lost in the first round must have started second, since otherwise they would have started first in their only series. So, every player who won in the first round must have started first. Similar reasoning works for the second round and so on, but then the tournament winner could never have started second.)

False.

Consider the case where in the first round, every player who started first lost.

If you start first in the first round and lose, then you started first in every series you played.

Oh, now I see what you're saying. Ok. Sure. But that doesn't invalidate anything else said.

[blueblimp]

Please, do not use the same number thoughout the tournament. That would mean that someone with a really low number would ALWAYS go first. It should be random every match.

For any way that first players are chosen and for any tournament results, it's always true that there is some player that starts first in every series he/she plays. Try it out.

(Proof idea: Suppose each player started second in some series. Then every player who lost in the first round must have started second, since otherwise they would have started first in their only series. So, every player who won in the first round must have started first. Similar reasoning works for the second round and so on, but then the tournament winner could never have started second.)

False.

Consider the case where in the first round, every player who started first lost.

If you start first in the first round and lose, then you started first in every series you played.

Oh, now I see what you're saying. Ok. Sure. But that doesn't invalidate anything else said.

Sure, it's just interesting that it's true.

[Turambar]

Bah. I hate the PCE and would much rather prefer it be banned if one player in a match requests so.

sure, but as we have said we can't enforce anything. ideally, the approach to this situation would be similar to what you see from other competitions with strong DIY or community based roots: the '

Some of my irl friends are in the tournament, they are ranked between level10-30. I suppose there would be no problems with me helping them when playing against high-level opponents? I mean, since you can't enforce any ruling against it...

[Beyond Awesome]

Bah. I hate the PCE and would much rather prefer it be banned if one player in a match requests so.

sure, but as we have said we can't enforce anything. ideally, the approach to this situation would be similar to what you see from other competitions with strong DIY or community based roots: the '

Some of my irl friends are in the tournament, they are ranked between level10-30. I suppose there would be no problems with me helping them when playing against high-level opponents? I mean, since you can't enforce any ruling against it...

You are right that it can't be enforced, but ethically, it is cheating, imo.