A problem that must be corrected.

[Adv1sor]

Easy enough to fix.

A tournament should be just that. A tournament. Not some giant league play.

When you lose a round you should be out.

Make the tournament two out of three, no time limit, single elimination rounds.

That would solve a lot of problems and the tournament would be over a lot quicker.

We should never have to look to the software to determine who won.

 

[Pablo]

yeah Adv1sor that will encourage every player to travel long distances to tournaments! great idea!

 

[Tego]

That would make this game almost as unsophisticated as sports.

 

[ninetales1234]

When you lose a round you should be out.

I don't like single elimination. If you get knocked out of a single elimination tournament, all that proves is that one person was able to beat you. I would reccomend you do some research on swiss and think about why we have it. Swiss is about finding the best player in the shortest amount of time. The TO gets you to play as many people as is practical. It is a mathematically sound idea, because the larger the sample size, the more accurate the results. The less people you play (or as in single elimination, sometimes, just one), the less reliable the results of the tournament being able to show how good a player you are.

In sinlge elimination, you're using an extremely small sample size to determine a player's ability. If you play someone in the first round that has an advantage over you (due to deck or playing style or whatever). How much does that really prove? All it proves is that one person was able to beat you. If the tournament was a 7-round swiss tournament, you could've gone 6-1, collecting 6 wins after your first round loss. But in single elimination, a first round loss always gets you last place (not literally, as there are tiebreakers in single elim to determine one's final ranking).

Everyone has an autoloss. Everyone has some person they can lose to, even Martin Moreno.

Larger sample size = more accurate results. If you want a format aimed at decreasing the workload for the TO, then single elimination (one out of one) is the way to go. If you want to find the best player in a short amount of time, then use swiss. Swiss has been around since 1895; It is not just some silly format invented by POP (maybe age modified swiss, but not swiss). To say that it is league play is extreme. I take swiss tournaments seriously, as should everyone else. There are strict guidlines to abide by when organizing a swiss tournament, so this is not league play at all.

And as others on this forum have pointed out, who wants to drive 3 hours to play in an event just to lose first round. This is sure to not make -10 parents happy!

We should never have to look to the software to determine who won.

This is something I don't like. Some people talk about swiss tournament results as though they are random or the computer "determined the winner". The computer doesn't determine anything. All it does is radomize parings and calculate tiebreakers. It's just doing math.

Tardiness
OP Win%
OP OP Win%
Head to Head
Standing of Last Opponent

These things are not random. They are totally legitimate attributes of one's tournament record and are reliable in determining how good that player is (though a bit flawed, some seem to think, as far as OP Win% goes).

 

[SteveP]

I'll have to look into the TOM 'flaw' if indeed it calculates resistance for dropped opponents by giving them losses for the rounds they missed.

I really like the DCI Reporter way of doing resistance, as pointed out by NoPoke. DCI Reporter doesn't penalize your resistance if you end up playing a bad-record opponent, up to a point. The minimum resistance for any one opponent is 33%.

 

[NoPoke]

It will be interesting to see the outcome of the 15+ grinder at this years worlds. Assuming that resistance is a factor in who makes the cut that is. After last year I expect many more players will drop with three losses and go and do other stuff on the Friday... like eat!

 

[Muk Man]

Instead of having resistance as a tiebreaker you could use the POP rankings?

because of players like me who dont get to go to that many leagues, or dont play in every sanctioned tourney available. i can only play when there is an event close enough, and on a day off. now if you're in...lets say Richmond VA, there will be the pre-releases, leagues, modified tourneys, cities, states, and gyms, all in you city, you will attend 2x as many events that will effect you pop rankings. so to say to me at worlds well he wins the tie breaker because he lives in memphis and his pop rankng is better.

 

[The Gorn]

I don't want to get off-topic, but there is a similar "problem" when someone is DQ'd. (Dairy Queened!)

I know at US Nationals a 10- player went from 16th to 17th following the DQ of another player (and, as a result, missed the cut).

I'm not sure what I'm suggesting here (which is why I put "problem" in quotes), but it just seems "bad" that in DQ'ing someone, those players who lost to a DQ'd player benefits (and rightly so) AND those players who beat the DQ'd player are punished.

It just seems like there should be a better way of doing this (although I will freely admit that I don't know what that would be).

 

[SteveP]

It just seems like there should be a better way of doing this (although I will freely admit that I don't know what that would be).

Just randomly select those in the cutoff range to fill the remaining Top X spots. That eliminates any unfairness (perceived or not) in the way the tiebreakers are calculated.

Better? Probably not.

Unbiased? Absolutely.

 

[Muk Man]

or... those tied for the remaining spot play 3-4 round of swiss... they wont mind, i promise.
the grinder will be crazy this year, i cant lie im so scared of whats gonna happen this year...

 

[Robbgobb]

I personally think that for Premier Tournaments should track people who drop. I know there are people who drop after a couple of losses. Then I hear people complaining about all the dropping. If this really a problem track dropping and see why it is done. Most that I see drop is just because of record. I have seen a few leave because of other reasons. If dropping is a problem then set up where penalties can be assessed to ones who appear to drop too often.

This is one thing though where I only suggest because people say is a problem for the players who had battled them. I go to play even when I do bad (started Nationals 3-0 and lost the rest). I do understand that people should not be punished for winning which is the normally reason why their earlier opponents dropped.

 

[ninetales1234]

It just seems like there should be a better way of doing this (although I will freely admit that I don't know what that would be).

If OP Win% tiebreakers being unfair is the problem, then POP could just utilize different tiebreakers. I'm surprised noone's suggested implementing "Harkness" (however, it still has the potential to be "distorted" by drops.).

"Time of Loss" would not be affected by opponent drops. Neither would "Cumulative". However, the "Time of Loss" tiebreakers are usually able to be determined before the last round is complete, so I would prefer "Cumulative" (since no one will know its final effect until the last round is over) as the tiebreaker after "Tardiness" (instead of OP Win%).

I think we should use Cumulative. Reward players for playing tougher opponents earlier (by winning matches earlier). No potential to be distorted by opponent drops.

So, I propose- instead of:

1. Tardiness
2. OP Win%
3. OP OP Win%
4. Head to Head
5. Standing of Last Opponent

We should try this:

1. Tardiness
2. Cumulative
3. Time of Loss
4. Had to Head
5. Standing of Last Opponent

 

[Flaming_Spinach]

We could always go with Sdrawkcabs suggestion and take an Exact Cut.

After Swiss, take everyone with a 66.6% winning average or better, those people make the cut. No one gets cut on resistance. Resistance is used for seedings only.


And yes, I do realize that this would make events longer, but I already touched on a way to fix that problem in a previous thread.

 

[DukeFireBird]

grinders are gonna be the same as last year i think, when 8 players are undefeated, were done

 

[Flaming_Spinach]

grinders are gonna be the same as last year i think, when 8 players are undefeated, were done

Please read the FAQ before you make a statement like that.

 

[ninetales1234]

We could always go with Sdrawkcabs suggestion and take an Exact Cut

I think that takes little the fun out of the tournament. The suspense, those moments before the T8, T16, etc is posted, you wonder what will happen to you- so exciting! But really, I don't think the exact cut idea is all that bad.:thumb: I think, it may even be better than using tiebreakers to determine who gets in single elimination.

However, such an idea assumes that the only thing to consider is who will make the single elimination cut. Tiebreakers will still be needed to determine prizes. Also, keep in mind that not all tournaments have single elimination, so the need for tiebreakers (and, consequently, the need to make them fair) still exists.

As I always say, we could avoid a lot of trouble if we just got rid of the single elimination and just try to find out who is the best player (by using swiss), instead of who can go undefeated for the longest.

 

[pop_webmaster]

First off, please know that POP has been discussing this issue internally since TOM first came out. We are doing a lot of research on the issue, but we really try hard not to change proceedure and policy mid-tournament season.

Regarding the suggestion below, I first want to say that it is a good suggestion and I welcome many more from you all. However, please think about this. While it is true that the Cumulative and Time of Loss tie breakers are drop-proof, they are also equal or shared between many more players. This means that more and more players, when compared to each other, will make it to the fifth tie breaker. You talk about how drops are totally out of the player's control (the non-dropping player), well the standing of the last opponent is even more so. Also, I haven't done the math, but I believe that the risk of having a head-to-head cycle is much higher.

So, I propose- instead of:

1. Tardiness
2. OP Win%
3. OP OP Win%
4. Head to Head
5. Standing of Last Opponent

We should try this:

1. Tardiness
2. Cumulative
3. Time of Loss
4. Head to Head
5. Standing of Last Opponent

That said, I do think you are on the right track. No decision has been made by POP, but we do take your well-thought out suggestions seriously. We recognise that we are trying to quantify something that can never be 100% quantitied, a player's skill, and we will make the best effort possible to provide the best software we can to make this program the best it can be.

Thanks,
Eric

 

[ultimatedra]

i hate it when ppl drop cause their doing so bad.they dont ever think how it could affect a person they lost to.they just think of themselves..

WRONG!!!!

At Worlds 05, my friend david (windjet) went into the grinder and lost the first round...the guy was a poor winner so david dropped late that tourny thinking to skrew the guy in standings.

poeple do think about that when dropping sometimes. .. . and david (and other smart players) do think about how hard they will skrew their opponents when they drop.

~Professor Ultimatedra~

 

[Flareon]

I wish we could have Playoffs for tied people. Single Elim until on eneough are left to squeeze in. It's a time issue, of course. Possibly Sudden Death? Or less Prizes?

 

[ninetales1234]

@POPwebmaster: Now, I know that POP tournaments aren't the only tournaments that use "Solkoff" or "OP Win%" to break ties. So, have you done any research into how FIDE or USCF handle drops with such a tiebreaker?

While it is true that the Cumulative and Time of Loss tie breakers are drop-proof, they are also equal or shared between many more players.

True. There is less oppurtunity for variation in those tiebreakers. With Op Win%, there are so many different calcuations involved (compared to Cumulative), that the numbers you come up with have a lot more variety. Some of the OP Win% have four digits in them. But Cumulative tiebreakers have a little less variety. They can be integers between "0" and "X * (X/2 + .5)" , where "X" is the number of rounds. EDIT: The more rounds a tournament has, the greater the set of possible integers for this tiebreaker.

But I still stand by what I think about Cumulative. Perhaps we could still use OP Win% and OP OP Win%- only make it one of the last tiebreakers, only if Cumulative, Time of Loss and Head to Head cannot resolve a tie.

Also, I haven't done the math, but I believe that the risk of having a head-to-head cycle is much higher.

What do you mean? Could you explain?

That said, I do think you are on the right track. No decision has been made by POP, but we do take your well-thought out suggestions seriously.

Yay! I'm smart!:clap:

We recognise that we are trying to quantify something that can never be 100% quantitied, a player's skill

Absolutely.
Keep up the good work. thanx.:thumb: