maybe i’m just befuddled, but my point is if an expert and an under 1800 player, for instance, tie for first, with prizes of 200 for first and 100 for top under 1800, the expert gets 150 for the two-way tie as opposed to 100 as one would expect if another >1800 tied for first. is my thinking that convoluted?
thanks, …scot…
If you have only those two prizes, $200- 1st, $100-under 1800.
To make reference easier, lets assume a 2100 and 1700
There are a few ways to consider this:
A: First place is clearly BETTER than a class prize, as shown in #2. Therefore (assuming a well-designed tournament) a class prize winner should in the same group as a first prize winner should not exceed what is given for first prize, but should do as well.
So start the 2100 at $200 and the 1700 at $100, and shift dollars from one to the other until the prizes are equal. This requires shifting $50 of the first place money to the $1700, and they each end up with $150.
B: An alternative and equivalent approach, is to simply add one prize per player to the pool and divide by the number of players: ($200 + $100)/2=$150
Now, do the same thing if a 1700 and 1701 in place of the 2100:
Then they each split $200 50/50 and they each split $100 50/50 for $150 each or
Award one $200, one $100, and subtract from one and give to the other until they are equal or…
Add 2 prizes and divide by 2.
All the same.
Bad cases make bad law?? I assume that you think that $100 for the expert and $200 for the U1800 is “correct”? (They split the $200 and the U1800 gets the $100 under prize). That’s the only other systematic calculation that I can even think of in this case. Suppose instead that it’s $150 1st and $100 U1800. If they split the first and the second player gets the full $100, then the U1800 gets $175 which is more than 1st prize solo. That same “method” would have the U1800 winning the full prize fund if he happened to finish 1st alone. (Try that sometime and see if anyone comes back to your tournaments).
Any prize fund which has prizes for which not all players are eligible will have certain possible results which are difficult to handle intuitively. The point of having a set process is to remove discretion from the TD to decide what’s “fair”. The divide and share method generally produces a reasonable result, and never produces a result which is clearly wrong (such as one player winning more than 1st prize).
Your point about the value of having a set process that doesn’t allow the TD to award prizes ad hoc is spot on. As a practical matter, that decides the question in your favor.
But I think you are reaching too far when you claim that this is the only “fair” result.
Arguments about “fairness” are often confused because fairness is not a unitary concept. In the context of tournament design, I think it has three elements:
(A) fairness is doing what people expect will be done;
(B) fairness is treating everyone equally; and
(C) fairness is giving the rewards to the player who performed best.
Fairness (A) depends on what people expect. This thread has shown to my satisfaction that at chess tournaments, the divide and share method is what’s expected. So in the sense of fairness (A), that’s the fair result. Your intuition that it would be clearly unfair for anyone to win more than first prize is an element of your expectation, and I won’t argue with it, expect to say that what seems obvious to you is not at all obvious to me (or to Scot). And I think it’s worth noting that outside the world of chess divide and share would likely be considered distinctly unfair.
Fairness (B) might lend some support to divide and share, as it requires an equal share for players who are tied. But no one’s idea of fairness would rely exclusively on fairness (B) – the perfect fairness (B) result is to divide the prize fund equally among all entrants. So if performance is to mean anything, fairness (B) always must be mixed with with fairness (C) to some extent.
There are elements of USCF practice that show little regard for fairness (B) – seeding is inherently in derogation of fairness (B), and the “Dutch” pairing method in use by USCF is a particularly strong form of seeding. You need to be pretty heavily invested in the current system to be comfortable with the idea that everyone with a rating below the median for the tourney is probably going to get clobbered in the first round. In many other sorts of tourneys, manipulating the pairings in this way would have players up in arms.
Fairness (C) is, in this matter, ambiguous. It depends on what you mean by “best performance”. In one sense, two players who are tied performed the same, and should be rewarded equally. But in another sense, the 1700 player substantially outperformed the expert because the 1700 player’s performance relative to expectation was superior. At least as I see it, the very existence of an U1800 prize suggests that the tournament does care about rewarding performance relative to expectation. I think it’s reasonable for a U1800 player to look at the U1800 prize and think, “that money is set aside for people like me”.
Two comments: seeding is much more inherently unfair if it is a knockout event. Since Swiss System events have each player play the same number of rounds (or at least the opportunity to score the same number of points) it is much less unfair. We generally have no concept of “SemiFinals” or “Finals”.
Second, if we offer say, a $500 prize for U1800, our way of paying prizes guarantees that at least $500 goes to players rated under 1800 (with the provision that there are at least as many players rated under 1800 as there are prizes. If the U1800 prizes are $300-150-50 and there are only two players rated under 1800, players rated under 1800 may only win $450).
Alex Relyea
Actually, similar systems are in use pretty regularly outside the chess world. Think of the NFL or NBA or MLB playoffs, or any pro tennis tournament. For the first few rounds, top seeds are playing bottom seeds, the intent being that the top seeds survive to play each other toward the end. The Swiss (not Dutch) System of chess pairings has a similar goal. You postpone the “dream pairing” to the later stages of the tournament. And you might be surprised at how many first-round games are not one-sided. In a recent tournament that I directed, 3 first-round games went so long that I decided to adjourn them so we could get the next round started on time. In all 3 games, there was still reasonable doubt about the outcome.
The overriding concept is one prize per player. The U1800 player can insist on getting that prize all to himself (and this has actually happened several times at tournaments that I’ve directed), but then he gives up any share of the overall first (or second, or whatever prize he’s tied for) prize. If he wants to share the bigger prize, he has to also share his smaller prize (2 players, 2 prizes). No player ever gets a full prize to himself plus a share of another prize. And the system is deliberately set up so that the U1800 player’s share of the total is never less than his U1800 prize by itself would be. He’s always getting something extra (if not, he just takes his own prize and there’s no sharing at all).
Players should never be allowed to “choose” prizes like that. The TD should “choose” for them, as the choice will invariably determine who wins how much of the other prizes.
Alex Relyea
Actually, I don’t see a problem. In the cases I referred to, there was (and could be) no tie for the Class prize. The only choices were : (1) The Class prize goes into the pool with the other prizes and players (assuming the number of prizes is no greater than the number of players), and everything is divided equally; or (2) the Class player gets his Class prize and the remaining prizes are divided equally among the remaining players. If the result of #1 is that the Class player ends up with less than the amount of the Class prize, the rules dictate defaulting to #2 anyway. So by choosing #2, the Class player cannot deprive anyone else of any prize money. He either gets the same amount he would have gotten anyway, or he gets less than that and everyone else gets more (he’s basically setting the maximum amount he can win equal to the minimum amount that he’s guaranteed to win). In both cases, this was explained to the Class player and he chose #2 anyway (because he didn’t want to wait around for the full determination of the split). But no-one else would have any reason to complain. If he’s willing to settle for a potentially smaller prize (not only willing, but insisting on it), I’m not going to argue with him. He is the only one who can possible lose out in that scenario, and he agreed to it with full knowledge of that.
Please review Rule 32B5.
Alex Relyea
The 2nd and 3rd scenarios provided in that rule are more complex than either of the cases I’m referring to (and in all 3 scenarios, I come to the same conclusion as the rulebook), but in none of them is my basic principle violated: the clear winner of a Class prize should never receive less from a distribution than what his Class prize alone would be. He is always guaranteed at least the amount of his Class prize, and I can’t conceive of anyone complaining if he agrees to accept no more than that.
Nevertheless, the rule does state exactly what you said in your previous post, and I will bear that in mind in the future. In the cases I referenced, though, I don’t see it so much as the player “choosing” a prize, as agreeing to accept the minimum prize that he’s guaranteed instead of waiting around to see if he might be eligible for more. This hurts nobody. We have also had cases of players “donating” their prizes to help defray expenses (when tournaments were poorly attended and bound to lose money). Would that also be a case of a player"choosing" his prize (in this case, no prize at all)? It seems a bit silly to force a player to wait around for more money when he’s happy with the smaller amount.
The general rationale for setting up and distributing prize funds has been to make sure that the lower rated players receive a fair share of the total prize fund. After all, the average players make up most of the entrants at a tournament. Without their entry fees, many tournaments would not even take place.
In many tournaments, the prize fund is distributed 50% to place prizes and 50% to class prizes. That is usual for a weekend tournament. In large class section tournaments that are trying to draw in GMs, IMs, and aspirants for titles, the top section is usually supported by the fees paid by all of the other sections. To get the high prizes in each class section, you have to pay a lot of money to enter and perform very well. Class C and B players very often make more in prize money than GMs in the class tournaments. Whether that is “fair” is largely irrelevant. Do not begrudge the average players their prizes. Organizers need them to participate.
The pooling of class prizes with place prizes happens when there are ties and the class player has had an amazing performance. In dividing the tied prizes in US events run on the Swiss System, look at the pooled money as one prize that is shared equally among the tied players. In cases where the class player would benefit from only getting the class prize, that is done so that he maxes out on prize money. In no case should be denied the correct distribution of money. It is bad practice for the class player to take a smaller share than he is entitled to. If he wants to give a part of the correct share back to help an organizer out of going into the red, that is okay, but he does not really have to do that. As an organizer, after I pay out the correct prizes, I don’t really care if someone decides to give his prize money or a portion of it to somebody else.
The distribution of prizes as shown in the Rulebook, as well as in general practice, is guided by being consistent in its methodology. If the distribution changes from event to event, players do not know what to expect. They will complain they are being cheated by the organizer. That is bad for business.
In my experience as a player, TD, and organizer, for players to win a prize, they generally have to have a performance that is 200+ points above their rating. In the big class section events, to get a top prize in a class section you may have to play 400+ above your rating. There are a lot of sharks in each section wanting to win prize money. If you want to know why there is rating manipulation or sandbagging going on, there is your answer. Playing at these higher performance levels on a consistent basis is difficult for a class player, but with one glorious weekend, a player who is stronger than his peer rating pals might win a 5 digit prize. On the top level or open section of such a tournament, you often see multiple players tied for first place. Usually there are about 35 or more sharks fighting to get a share of only 10 place prizes. The lower place prizes are almost embarrassingly small. In the feeding frenzy, 2200 to 2300 rated players have to fight extra hard for their relatively small class prize compared to what their lower rated friends are playing for. If they do well, they sometimes have to share their prize money with higher rated players if they tie for place money. It is considered okay as long as these master class players max out their prize potential. All too often, they go home with nothing. Many GMs go home with nothing, too.
That’s a very good point.
The best seeding practices I know of are the ones used in ATP and WTA tennis, where high-stakes knockout events are the main form of competition. In tennis, usually only a quarter of the field is seeded, and the seeding is tiered.
It’s not clear to me why chess tourneys need to be seeded. But given that they are, the Dutch method is probably better than conventional seeding (e.g., 1 v 16, 2 v 15, … 8 v 9 for a 16-player event).
well, you would have assumed incorrectly. i’m just playing devil’s advocate here. and the splitting of the prizes IS probably the best way to do it. even though it does seem a bit unfair that a player rated above the 1800 limit would receive some of the prize money that was supposedly allocated to that subset of players.
cheers, …scot…
Scot - it was shown in your example above that the player rated above 1800 DOESN’T receive some of the prize money allocated to lower players.
The combine and split method will only include under prizes when the under prize is less than the split.
Example one
2nd=$300, 3rd=$200, U1800=$100, four way split with one U1800 player. The split gives $150, which is more than the U1800 prize. Although not an entirely accurate wording, you can look at it mathematically as the U1800 player trading 3/4 of the value of the U1800 prize ($75) to get 1/4 of the value of the combined 2nd and 3rd prize ($125) for a net gain of $50. The U1800 player receives $150 (more than the U1800 prize) while the other three receive $450 (less than second and third). If there was one less over-1800 player in the split then they would have all received $200.
The alternative would be for the U1800 player to receive only $100 so that the other players could each get $166.67 (three others getting $16.67 more than when including the U1800) or $250 (two others getting $50 more than when including the U1800).
I think people get hung up on the phrase (“of the value of”) and think that some of the over 1800 players are being given a piece of the U1800 prize and thus come out better than if the U1800 player was not in the tie (not accurate, if the U1800 had scored less and not been in the tie then the over-1800 players would have received more).
Example two
2nd=$300, 3rd=$200, U1800=$100, fifteen way split with two U1800 players. The split gives $40, which is less than a two-way split of the U1800 prize. Thus the U1800 players simply split the full $100 ($50 each) and let the other thirteen players split the remaining $500 ($38.46 each).
I remember hearing a (possibly apocryphal) story about an long-ago expert doing well enough to be in a somewhat large tie for one of the lesser place prizes at the US Open, with the result that the prizes were totaled and then divided between the tied players. That meant that the expert received less than he would have received if some of the players he tied with had done better and left no place prizes available to be put into the pot.
The story was cited as the reason for adding the qualifier that if the class player(s) end up better by not participating in the tie for open prizes then the class player(s) and the class prize(s) are carved out and the split is only done for the other players.
PS There was a case a little while back kind of similar (dollar amounts and number of tied players may not match but the gist is there) to having a master and two experts tie for third ($300), fourth ($200) and top expert ($500) with a split of just the expert prize giving the two experts $250 each while a split of all three prizes would give everybody $333.33 each. That resulted in additional wording that a person only eligible for non-class prizes could not receive more from a tie than would have been received by being the sole recipient of the top prize still available to the player (the distribution actually done, and the one thus codified in the rules, was that the master took the $300 third place prize while the two experts split fourth and top expert for $350 each).
Unfortunately, the current rule book language is still flawed. While most experienced TD’s know what is meant, the language that seems to single out individual players for exceptions runs into problems when there are multiple ties for multiple categories of prizes. See
viewtopic.php?p=311884#p311884
for a more technically correct description. (I thought Bob Messenger was going to submit that as a rule change several years ago, but apparently not).
The current rule is
Where that could run into an interpretive problem is if you add one Master and one more place prize to the situation cited by Jeff. Make it third ($300), fourth ($200), fifth ($100) and top expert ($500). The problem with the first sentence is that the fifth prize is not (strictly speaking) a prize for which the two masters are “ineligible”. Taken overly literally (which was the problem), that means that the expert “hold-out” would only split the $500 (for $250 each). The four way split of all prizes is $1100/4=$275 so that would be preferred. Note that the masters then get less than $300, so the second sentence has no effect.
Note that that requires an ultra-technical interpretation of “ineligible”—ex ante, the masters are eligible for 5th prize, but once they are given shares of 3rd and 4th, they aren’t. The correct way to look at this is to hold the experts out, let the masters max out first (($300+$200)/2), then let the experts max out with what is left (($100+$500)/2). Since $300>$250, the experts are better off that way. If it were a $300 expert prize instead, then the hold-out calculation would give $200<$250, so you would go with the even four-way split.
This is a reason why the method of thinking of taking dollars from the “upper prize” and shifting it to the “lower prize” until they are equal can be a useful way to consider it; there are no upper dollars to take.
I didn’t see anything in the OP about clear winner prizes. Why is 1st place not given out? Sorry for the lateness of seeing this.
From the OP:
This is admittedly a bizarre wording, as taken literally, it means that there is no 1st place if there is no clear winner. Perhaps the intent was to “clarify” that ties for 1st didn’t all get $100, but that’s an odd way to state that.