A new structure for 1000+ entrant tournaments

[t_perkin]

I think I am not alone in believing that the current tournament structure does not allow the cream to float to the top in 1000+ entrant tournaments.

The question I want to ask is: is there a better tournament format that could allow 1000+ events to take place and still retain a large element of skill? Do other tournament formats already exist?

Here is my suggestion:

All players sit at 10 player tables
10 players sit at a single table for 2hrs and the top 2 money holders are taken to the next round, blinds increasing - but relatively slowly. Tables do not get amalgamated as people get knocked out.
If the table still has more than 2 players after 2 hours then tough luck, the top 2 go through, other places get eliminated.
Repeat the process with the players which have made it through. either with the players continuing with the chips they had or possibly with everyone starting from scratch again in the next round, or possibly somewhere between the two.

This way you can get a 1000 players down to 10 in 6 hours play exactly. However you still get to sit at a table for 2 hours with the same players rather than constantly changing tables and without the constant pressure of raising blinds.

There are lots of problems with this format (everyone grabbing for chips at the end of each 2 hours etc.). But I just wanted to see if people had better ideas for tournament structures.

Tim

[Greg (FossilMan)]

Why do you believe that the proposed format will allow the more skillful players to do better than a format that follows the traditional pattern and lasts about the same amount of total time?

In my view, all that matters is how many players started and how long the tourney lasts. The longer it takes, the more skill is rewarded (on average). The shorter the format, the less skill is rewarded. I really doubt there is any way around this.

Your suggestion would allow somebody who is better at short-handed play to do better than somebody who's talent is full-table play, since they would spend more time playing short-handed. However, I don't think this structure would reward overall poker skill more than the traditional format in the same time interval, it would just reward certain aspects of overall poker skill more than the traditional format.

Later, Greg Raymer (FossilMan)

[JDO]

I would be afraid that the real effect would be opposite from the goal. I think your system may punish tight-aggresive play; where variance is small over the tourney.

Players that tight aggresive players describe as bad players are often discribed with words like Maniac, loose and over aggresive. These guys tend to have higher variance than tight aggresive players. They would therefore have a high variance. Which means that in a very short term, place top 2 at your table or go home, a player with a high variance (assuming he is hitting) might have a better chance of moving on than he normally would. Yeah, he'll bust out early, but he'll get a huge stack early sometimes too. So, I think your system selects for the loose, aggresive players succeeding more often and I think you would see more "bad" players move on into later rounds than under the current system. Just my thoughts.

[Bozeman]

This might allow for slightly more skill if the stack sizes are all reset at each change.

Also, less table moving might have a small effect.

[t_perkin]

The thing about the structure I proposed is that is scales much better. You can have a tournament with 10,000 players and be down to two tables in 8 hours. Yes maybe we will never see tourns with 10,000 players in, but then who thought we would see tourns with 1500 players in?
A tourn with 5000 players is not too far in the distant future, and my struture would allow this to get to the final table in 8 hours. The existing structure would take at least 25-30hours min.

I am suggesting that the blinds would increase very slowly in each 2hr period, which would allow people to sit a bit tighter. Yes you may be right, I can´t think how to avoid the "chip grab" at the end of each 2 hour session. I think it would be all about getting the blinds to increase at the right speed.

Anyway, this was meant as just an example of a different structure. What I was really pointing out was that the existing structure was not designed to cope with 1000 players let alone 2000+.

As you acknowledge, different structures do affect the skill level of the tournament. So my question is, what *is* the most skillful structure for these massive tournaments? Is it really the structure that is in use today?

Tim

[Greg (FossilMan)]

The most skillful structure is the one that gives the players the most playing time. The more time over which you are measuring somebody's poker performance, the less luck is a factor. And that's about it for reducing luck.

The only reason a traditional format would be any slower than your format (with the luck factor being even), is that the process of breaking tables and moving players would cause players to NOT get to play as many hands in the same time period, due to the time spent moving around. Other than that physical limitation, it's all just a matter of how fast the blinds increase (which is directly proportional to how long the tourney will last, which is directly proportional to the luck factor).

I very much doubt there will ever be a LIVE 10,000 player tournament with any structure, traditional or otherwise. So the concern is probably academic. If I'm wrong, I hate to imagine the chaos at the live game sign-up board as people bust out of the tourney. ;-)

Later, Greg Raymer (FossilMan)

[t_perkin]

I am afriad that I have to disagree.

A tournament is effectively a sorting problem. The aim being to sort the players into the correct order according to ability.

Now to explain my reasoning I am going to simplify by saying that if player A “is better than” player B and player B “is better than” player C then player A “is better than” player C.
For those of you mathematically minded I am saying that the "is better than" relation is a transitive relation,
I am also assuming that:
If A "is better than" B and B “is better than” A then A has equal ability to B.
For those of you who are mathematically minded I am saying that the "is better than" relation is antisymmetric
I will finally make the assumption that if A is better than A then A has equal ability to A. This is a trivial result.
For those of you who are mathematically minded I am saying that the "is better than" relation is reflexive.

The transitive property is the most controversial, however this is an assumption that is made in the structure of all tournaments and “ladder” based competitions including the traditional poker tournament.
The transitive property can be avoided by having competitions such as
leagues, and *partially* avoided by using seeding and “group stages” (which are “sub” leagues) in tournaments.

The transitive property would be impossible to elliminate in a poker tournament, but it could be alleviated by the use of “group satges” etc. . The traditional poker turnament has no system of alleviating the transitive property. So the transitive property must be accepted if an argument is to be made regarding the validity of the traditional poker tournament in comparison to any other system.

Having made these three assumptions we can now say that the “is better than” relation forms a partial ordering (reflexive, antisymmetric and transitive relation). This means that we are able to theoretically order all players into order of their ability (although players can be considered to have equal ability).

Now that we have a partial ordering we have reduced our problem to a sorting problem. I am no specialist in sorting algorithms, and this problem has some very interesting quirks in terms of which sorting algorithm would be most suitable. But I do know and any mathematician will tell you that sorting algorithms do not all have the same level of effiecency (in terms of time). Therefore there are objectively better and worse ways to run a tournament, even with the same level of “accurateness”.

This is not an argument that my new tournament structure is "better" than the existing one. Only that you cannot say that all tournament structures are equal if they are of equal length.


Tim

[gavrilo]

There is a Tournament like this where I live.
I don't know if it's the top two or just the top player at the table advances to the next round and each round is like an hour or so.
The problem with this is that the last two hands of the round are crazy, everybody just goes all -in, if you don't win one of the last two hands most times, you are out.
It's just a big crapshoot.

[Paul2432]

[ QUOTE ]
A tournament is effectively a sorting problem. The aim being to sort the players into the correct order according to ability.

[/ QUOTE ]

Well I disagree with this. I don't know anyone that runs a tournament for this purpose. The tournament organizers run the tournament to make money. The players enter the tournament to win money and possibly have fun. I know when I enter a tournament my goal is not to determine my ranking amongst my fellow players but to win the tournament.

Does anyone here enter a tournament not to win, but to know how they rank amongst the other players?

If I may read between the lines, my guess is you have been getting poor results in large tournaments and you think it has something to do with the structure.

Paul

[t_perkin]

Ok so you enter a tournament to win money and not to find your ranking, and let us assume that so does everyone else. (yes there may be people with no interest in winning money but they do at least wish to win the tournament - another assumption, but a small one).

So you want to win money - which means you must believe that you will get a +ve return on your investment in entering the tournament. i.e. you will make more money than you put in on average.

And what makes you believe that you are going to make a profit? well it must be the fact that you are more likely to finish in the money than your opponents.

And why are you more likely to finish there than your opponents? because you are a *better* player than them and so you will win more hands and more tournament chips and get knocked out of the tournament later. the better the player you are the longer it is before you get knocked out, and longer you last in the tournament and the more money you make.

So the whole point of a tournament is to reward the best players. If the tournament didn´t reward the best players then it would just be a lottery like any other.

In order for the tournament to reward the best players it must order them from best to worst. It is therefore a sorting problem. It starts with an unsorted list of players. and it trys to end up with a sorted list of players, best to worst.
The better it is at doing this the more a skillful player will be rewarded.

Tim

p.s.:
If I may read between the lines, my guess is you have been getting poor results in large tournaments and you think it has something to do with the structure.

I have only ever played 4 multi table tournaments in my entire life. I have finished 16th, 29th, 50 something and 100 and something. They have all been less than $20 buyin , 350+ entrant affairs. My ability is irrelevant I am not posting to whine.

I am an (admittedly pretty inexperienced) Computer Science Lecturer and my areas of interest are compilers and directed graph theory.
I am posting because I feel it would be interesting to explore the idea of different tournament structures (which is related to graph theory). I´m sorry if I sound like I am complaining, I am really not. Maybe I should be posting in "General Theory" rather than "Tournaments"?

[Copernicus]

I would also disagree with the premise that it is essentially a sorting problem that can then be related to sorting algorithms. Sorting algorithms deal with known attributes and constraints, they do not deal with discovery of the differentiation in levels of the attributes.

Your concerns about tournament length are more akin to problems in parallel vs sequential processing, and placing a time constraint on each processor (table) results in a loss of information (the all-in fest when close to the time limit replaces skill with luck, or at least emphasizes a different set of skills).

As others have pointed out, the problem eventually comes back to the relationship between time and skill. The most obvious example of where skill is de-emphasized in the interest of time is the Party SnG relationship between initial stacks and the blind structure. They make a fixed fee per tournament player, and can run more efficiently (ie maximize the return on investment in computer hardware) the shorter those tournaments are. Thus they have a different structure for lower buy-ins (where presumably the skill levels do not justify lengthening the tournament) than at higher buy-ins (where presumably there is more skill differentiation and awareness on the part of the better player of how the structure impacts his chances of winning).

While there is still a loss of information by accelerating blind increases or shortening stacks that seems to me to be less artificial than a time limit on a table.

With regard to distributing players periodically to balance the number of players at each table vs playing each one to a winner or two (without a time limit), there are at least two effects. One is the noted shift in emphasis between full table and shorthanded play skills. Perhaps that is desirable. The current structure only allows for shorthanded skills to be demonstrated once, by the survivors to and of the final table, instead of demanding them from 1/3 to 1/2 of the players in the tournament (depending on where you being to call a table shorthanded) The other effect is the desired reduction in time to eliminate the same number of players. There is some reduction in time because shorthanded play moves more quickly than full tables (eg two tables of 5 would get down to 2 players somewhat more quickly than one table of 10). However, even those savings are reduced by the need for table winnner/winners to wait for a sufficient number of other winners to start up the next table....ie processing is constrained by the slowest processors. The net effect may be minimal.

The concern that the numbers of players will grow to a point where the time to complete them is inordinate can be addressed in another way...increase the number of different buy-ins (ie add higher limit tournaments) and the prizes accordingly. That in itself begins the differentiation process in what is likely to be the most efficient manner. Each player brings a lifetime of play and self-evaluation to the determination of what level he should buy into, and that level of processing is instantaneous. While that does introduce an element of wealth/bankroll to the initial differentiation, presumably as stakes go up even the very flush but unskilled player will realize they are dead money at a given level.

The result is that you have more massively parallel processing. Each table is now less constrained since with fewer other tables that need to be completed (in the winner only structure) before a round can be finished and the next commenced, the probability of a slow table constraining the others is reducd.

[Greg (FossilMan)]

Sure, I agree it's a sorting problem. And I agree that factors other than time do matter.

However, time is the overriding factor.

If we invent a structure, any structure, that results in a given event lasting an average of 5 hours, there is more luck in that event than if we made a new structure, any new structure, that gave the same number of players an average of 7 hours of playing time.

There are exceptions of course, but I firmly believe that time is by far the overriding factor in determining how much a given tourney structure rewards skill.

Later, Greg Raymer (FossilMan)

[Lottery Larry]

i am not a tournament pro or anything but have you not just reduced your big tournament to a crapshoot 2-hour satellite with multiple tables?

[Eric P]

your tournament structure seems to be dealing with tournament length as the main problem. Which is going the complete opposite argument that current (long, WSOP 5+ days) tournaments don't allow the betters to win. You are contradicting yourself

[Copernicus]

Great analogy LL

[t_perkin]

I think the new structure I suggested is less skillful than the existing system. Thank you everyone for your thoughts on that.

Are there other systems that could be played over the *same* amount of time as a traditional tournament which offer *less* variance?

I am not going to post anymore on this, otherwise I think people are going to get annoyed at me /images/graemlins/frown.gif /images/graemlins/smile.gif

Thanks everyone for your thoughts.

Tim

[Ignatius]

I think I am not alone in believing that the current tournament structure does not allow the cream to float to the top in 1000+ entrant tournaments.
.
Your project is somewhat contratictory: The very reason that there exist tourneys with 1000+ entrants is exactly because the "cream" does not always float to the top. If only the top players have a chance to win, only the top players will play and there will be no dead money to play for. If you want this, you better take up chess. At least in chess, there is some sponsor money to be won.
.
That being said, there is also a fundamental problem with your format - or any other format where chips get devalued for most players at some point during the tourney: Collusion. In your case, simply team up with some buddies and contract that at some point near the end of the level, they will give their stack to the chip leader amongst them. Playing this way, 2 buddies will gain more than their "fair share" of 2/5 places as they can combine their stacks whenever neither of them qualifies to move on, while their fair playing opponents don't have this option.