Prize distribution revisited

Running Chess Tournaments
1

Had an interesting place/class prize distribution problem this weekend. Rob Getty had a topic in July 07 with a similar theme.

My event had 4 place prizes with the 4th = $100. It also had several class prizes of $80 - A, B, C, D, & U1200. 1st went to the clear 4-0. There was a 4 way tie at 3.5. SwissSys’ prize calculation module awarded the 2nd, 3rd, 4th and B to this group. I think it used the following logic to assign awarded prizes and then summed and divided the prizes equally.

List the players in rating order and award prizes based on eligibility.
Abe rated 2208 got 2nd
Bill rated 1943 got 3rd
Chas rated 1798 got 4th
Doug rated 1792 got B

I agree with several TDs who commented in the earlier topic that with equal money, you still award by considering that A is better than B and the A prize, not the B prize should go into the pool. Do you agree?

Now the tricky part if you agree the B prize goes into the pool. If Bill was the only class A player in the tournament, would you still include the B in the pool and pocket the A money? If you included the A prize in the pool, what if Chas and Doug were the only Bs in the event?

FYI, I paid the B prize to the 3-1 group. Also, in past events where I had to include only one of 2 class prizes, I included half of each class prize so the players in the next score group got some money too. I like this when the class prizes are equal because the players in the higher group get the same amount but it spreads more money among other players in a lower group.

Regards, Ernie

2

This depends entirely on how you defined the prizes in the advance publicity. If you listed the prizes as “Class A” and Class B," the distribution given by Swiss-Sys is correct – A-players are not eligible for the Class B prize, 4th is larger than Class A, so the B prize goes into the pool and the A prize is not awarded (no one eligible). If, on the the hand, you advertised the prizes as “Under 2000” and “Under 1800,” the U2000 prize should have been pooled first, and the U1800 would go down to the next scoregroup.

Later: Looking at it some more, I’m no longer quite sure. Chesschick’s argument has some merit. Also, as a practical matter, pulling the A prize into the pool has the advantage of awarding all the prizes, so it probably deserves a small edge. There are a number of possible comments about how the problem was created by the prize structure (e.g., if they were “under” instead of “class,” there wouldn’t be any question about it), but that doesn’t really help much with the specific problem.

3

However, Bill is eligible for A and 3rd (or 4th place), and A and B are the same. So, if you put the B players in the artificial 3rd and 4th places, and then the A player would be bringing his prize into the pool. So, why would you pool the B prize over the A?

I think I would have pooled the higher prize as well.

Maybe this is one situation where the fractional method might be more fair? You take 1/2 of the A and 1/2 of the B and pool it, then give the other half to the next player(s) eligible in the class. I have never used this method though.

4

I remember from a tournament in Arizona that the TD was faced with a similar dilemma. I wasn’t the TD, but it interested me to find out how some of my NTD friends would have handled it. After talking to them I came to the conclusion that the “best solution” would be to take 50% from both the A and B prizes to award it to the 3.5 score group and to award determine the remaining A and B prizes (separately) as usual.

It looks to me that the same method could be used here.

My above recommendation takes the “choice” our of my hands… in both cases I would be pocketing $40, which might look bad. But it would look worse if I made a decision that pocketed $80 instead. I like my solution because I wouldn’t have to “justify” it for situations were reversed in a future event. (“Last time you took the higher/lower class!”)

Although if it were that specific of a case, I would probably choose the situation that didn’t have me (the tournament) pocketing anything. If all the B-players are in the score group it partially makes sense to make a “one-time” decision.

I agree.

  • Enrique
5

Is there any compelling reason to even allow “A” and “B” prizes, as opposed to U2000 and U1800 (other than being able to describe them with fewer characters)? If you use under prizes rather than strict classes, the prize allocation calculation is unambiguous, and you avoid the possibility of a lower rated player with a higher score getting less money than a higher rated player with a lower score.

6

I will admit that I sometimes get scared by how several experienced TD’s can look at the same prize distribution problem and come up with different solutions. This is one area that there should be no ambiguity. When people expect a certain prize breakout, and a different (yet arguably valid) method is used that results in less or no money for them, it makes for some really bad situations.

7

Yes. Advertising a “Class A” prize in an area where there are very few Class A players may induce more Class A players to enter your event.

The bottom class prize should always be an “Under” prize - but I have seen many events where it makes sense to use strict Class prizes for the higher classes.

8

Stripping the example to its essentials:

Prizes:
4th place = $100
A = $80
B = $80

Results:
1 player at 4-0
4 players at 3.5:
2208
1943
1798
1792

Everybody seems to agree that $80 should be thrown into the pool for 2nd through 5th, thus combining 2nd, 3rd, 4th, and one of the class prizes, and dividing the sum 4 ways equally.

Some argue that the $80 that is thrown in should be the A prize, others say the B prize, still others prefer to throw in half of each. It makes no difference to these four players, but it makes a huge difference to the players next in line (A and B players with 3-1 scores).

What happens if the prize structure is slightly different:
4th place = $100
A = $80
B = $79

Now, according to the conventional logic, $80 must be thrown in, since it is the larger prize. This appears to mean that the A prize is thrown in, while the B prize goes to the next player(s) in line (presumably a B player with 3-1).

Now let’s change it again:
4th place = $100
A = $79
B = $80

(Why would the B prize be larger than the A prize? For the same reason the organizer awarded class prizes rather than “under” prizes to begin with – because he expected more B players than A players.)

Once again, conventional logic says to throw in $80, the larger of the two class prizes. This time, we must apparently throw in the B prize to accomplish this, leaving the A prize to the player(s) next in line, i.e. an A player with 3-1.

This conventional logic, however, creates a Stupid Discontinuity at the point where the prizes cross over. For example, as the two class prizes cross over from A=$80, B=$79 to A=$79, B=$80, the result changes from $79 going to a B player to $79 going to an A player. A mere $2 difference in the prize structure has resulted in a $79 difference to each of two players.

If, in the case of equal prizes, A=$80, B=$80, we throw in $40 of each prize, now we have two Stupid Discontinuities:

Prizes:   Result:
$80,$79   $80 to the place pool, $79 to a B player at 3-1
$80,$80   $80 to the place pool, $40 to each of two players (A and B) at 3-1
$79,$80   $80 to the place pool, $79 to an A player at 3-1

The flaw in all of this conventional reasoning is the assumption that, because $80 should be thrown into the pool, it needs to be that $80 (A or B prize as the case may be) that should be thrown in.

The solution is to apply the split philosophy ($40-$40 split) also to the case where the prizes are not equal. The $80 that is thrown in could be composed of parts of each prize. For example:

Prizes:   Result:
$80,$79   $80 to the place pool, $79 to players at 3-1 ($40 to an A, $39 to a B)
$80,$80   $80 to the place pool, $80 to players at 3-1 ($40 each to an A and a B)
$79,$80   $80 to the place pool, $79 to players at 3-1 ($39 to an A, $40 to a B)

The Stupid Discontinuity is gone. The prize-awarding algorithm is now a continuous function – small differences in the prize structure result in only small differences in the prizes awarded to each player.

Better, no?

Bill Smythe

9

Hi Bill:
Are you proposing this as the standard case? If yes - and I agree - then what should be the procedure if an old time TD wants to use the previous standard. I feel that there is still a dispute in the case being discussed over which prize, A or B, should come into the pool if equal value. We should resolve this dispute now to stand until or if the rule is changed. I vote the higher ranking which is first value (cash amount) and if that is tied, the prize with higher status X>A>B and so on. This would also apply to under prizes. UX > UA > UB. I also assume that if two under prizes could come in, we would divide those also, correct?
Regards, Ernie

Bill S said:
This conventional logic, however, creates a Stupid Discontinuity at the point where the prizes cross over. For example, as the two class prizes cross over from A=$80, B=$79 to A=$79, B=$80, the result changes from $79 going to a B player to $79 going to an A player. A mere $2 difference in the prize structure has resulted in a $79 difference to each of two players.

If, in the case of equal prizes, A=$80, B=$80, we throw in $40 of each prize, now we have two Stupid Discontinuities:

Prizes:   Result:
$80,$79   $80 to the place pool, $79 to a B player at 3-1
$80,$80   $80 to the place pool, $40 to each of two players (A and B) at 3-1
$79,$80   $80 to the place pool, $79 to an A player at 3-1

The flaw in all of this conventional reasoning is the assumption that, because $80 should be thrown into the pool, it needs to be that $80 (A or B prize as the case may be) that should be thrown in.

The solution is to apply the split philosophy ($40-$40 split) also to the case where the prizes are not equal. The $80 that is thrown in could be composed of parts of each prize. For example:

Prizes:   Result:
$80,$79   $80 to the place pool, $79 to players at 3-1 ($40 to an A, $39 to a B)
$80,$80   $80 to the place pool, $80 to players at 3-1 ($40 each to an A and a B)
$79,$80   $80 to the place pool, $79 to players at 3-1 ($39 to an A, $40 to a B)

The Stupid Discontinuity is gone. The prize-awarding algorithm is now a continuous function – small differences in the prize structure result in only small differences in the prizes awarded to each player.

Better, no?

Bill Smythe
[/quote]

10

The most important things in distributing prizes are to have clear, fixed rules in place and to follow them absolutely.

I don’t see any problem with the current rules. Leave them alone. If you’re going to use different rules for your own tournament, just make sure you clearly announce it in advance in all tournament ads & flyers.

I have pretty much decided not to play in a local organizer’s events just because once he didn’t follow the written rules and advertised prizes and ended up giving me $30 less as a prize than the rules called for (he “combined” a couple of class prizes after the start of the tournament and I ended up sharing a prize when I would have won one outright in the announced prizes).

11

The first blunder was by the organizer giving each class the same prize. By just changing the prizes by one dollar in each class, the prize distribution becomes much more automatic and definitive. Also using Under-xxxx can be very handy.

All the best, Joe Lux

12

No. Everyone eligible for a U1800 prize, say, is also eligible for a U2000 prize. The U2000 prize is more prestigious, so it goes into the pool.

Alex Relyea

13

Yes, we need a change, and this is my initial proposal. More discussion, however, is required first, to work out whatever kinks may still be in there.

Any time any new rule is enacted, any TD who ignores the new rule is subject to the “lynch procedure”: the players will eventually get on the TD’s case, and sooner or later the TD will change. That’s what happened, for example, in 1994 when “a delay clock is a preferred clock” superceded “black gets his choice” as the standard rule for determining which player’s clock to use.

I was hoping the conversation would focus on the details of how the rule should be changed, rather than on what to do until the rule is changed. But, sigh, I guess the latter topic is important, too. I suppose I agree that the Stupid Single Discontinuity (A is more prestigious than B) is preferable to the Stupid Double Discontinuity (divide it 50-50), although I’ll admit the latter has a lot of appeal. Meanwhile, the best interim work-around would be for organizers to make each class prize smaller than the one above it, to avoid the Stupid Discontinuity in the first place.

A Stupid Discontinuity is not a problem? I strongly disagree.

That’s true, of course, but organizers continue to make this blunder, so the rule needs improvement.

Apples and oranges. Ernie is talking about what to do if the rule is changed. You’re talking about what to do under the current rule.

Bill Smythe