ADM: Modify Chapter 12 of the US Chess Rulebook

Running Chess Tournaments
1

I am hoping a delegate will sponsor the following motion at this year’s US Open:

ADM: Modify the first paragraph of Chapter 12 of the US Chess rulebook, Round Robin Pairing Tables, as described below.

The paragraph currently reads: “The following pairing tables are used for round robin tournaments. The player with the first number in each pairing has the white pieces. Pairing numbers are assigned by lot at the beginning of the event, unlike Swiss tournaments in which pairing numbers are determined by ratings.”

Change this to: “The following pairing tables are used for round robin tournaments. The player with the first number in each pairing has the white pieces. Pairing numbers are assigned by lot at the beginning of the event except for three and four player round robins in which, like Swiss tournaments, pairing numbers are assigned in order of rating.”

Rationale: Part of the first paragraph of Chapter 12 of the US Chess rulebook is trying to explain how pairing numbers in round robin tournaments are determined but makes no mention that pairing numbers in three or four player round robins are assigned in order of rating as explained in rule 30G which states “Players’ numbers are assigned in order of rating, not randomly as in larger round robins.”

2

Is that even a “rule”?

I agree that this needs to be changed since there is no earthly reason not to make random assignment of pairing numbers in quads the norm (since that’s easiest for computers) and make it optional to use the manual system of rating-based assignments.

3

Just to clarify, are you saying random assignment of pairing numbers in quads should be the norm just because that’s easiest for computers?

4

It should be the norm, because it’s the norm for round robin tournaments in general.

5

If you want to propose a rules change ADM, you would do well to communicate with the chair of the Rules Committee.

6

Good luck with that. I tried to get the round robin tables changed about 9 years ago (US Open in Dallas), but was told I didn’t correct a serious enough proglem to examine further. rrpair.htm

7

No it is not a rule.

In fact I would agree that since it is not a rule the delegates have no authority to change it.

change 30G - now that is a rule the delegates can change. But chapter 12 is not a rule.

8

When we do Quad tournaments, the players in each section are not necessarily placed in rating order. When the same players show up all of the time it is better to mix it up so that they are not always playing the same color versus opponents. Other organizers do something similar in their Quad tournaments. Often enough, section order is based on when a player came in to register. The earlier you register, the better chance you may get White twice in the event.

9

I’m confused. Let me see if I understand this correctly.

Chapter 12 says that player numbers in round-robin events should be determined by lot, not by rating. This would include quads.

Rule 30G says that, in quads, player numbers should be determined by rating.

Micah Smith wants to change Chapter 12 to agree with 30G so that, in a 4-player round robin, player numbers should be determined by rating, but that in a larger round robin, player numbers should be determined by lot.

WinTDoan comes aboard and first says he agrees with Micah Smith, but then says that 30G should be changed rather than Chapter 12, so that player numbers should be determined by lot, regardless of the number of players.

Then tmagchesspgh says he assigns player numbers by lot in quads.

If you ask me, assigning player numbers by lot, across the board in round-robins of any size, makes the most sense, displays the most consistency, and avoids giving the top-rated player an extra white every time.

HOWEVER, in a quad, this goal can be achieved much more simply. Instead of tossing a coin for each player in each quad section, toss ONE coin. Heads, order the players in each quad in rating order, highest-rated first. Tails, order the players in each quad in rating order, lowest-rated first.

Bingo! That way, colors are randomized as much as they need to be, and the top-rated player (or any other player) has a 50-50 shot at getting an extra white or an extra black.

A slight variation (and just as good) would be: Heads, order the players highest-first in the top quad, lowest-first in the 2nd quad, highest-first in the 3rd quad, etc. Tails, just the opposite.

Maybe if a pretty system like the above were implemented, those who feel yechhy about randomizing (because of the multiple coin tosses) would feel a little more comfortable.

Bill Smythe

10

Huh? I don’t agree with Micah regarding what should be changed (and, as I point out with a rhetorical question, what’s cited isn’t even a rule that would be fixed by an ADM). I don’t have a problem with doing quads manually in rating order, I just think it’s silly to say that that’s the way to do quads.

11

I guess what confused me was that you said “I agree …” and then disagreed.

In any case, I certainly agree that it would be silly to make rating order the only way to do quads.

Bill Smythe

12

This is certainly an idea to consider

The first paragraph of chapter 12 re-states the rule on how pairing numbers are assigned in round robin tournaments but is incomplete and misleading since it doesn’t state that pairing numbers in quads are assigned by rating. The delegates approved changes to chapter 8, The USCF Rating System, last year to correct some mistakes and make some things more clear so there is precedence for the delegates to approve changes like this.

If the delegates don’t have the authority to change it, then who does?

That’s fine, although that is a variation on the current US Chess rule.

Under the current US Chess quad rules, colors are random in the last round so the top-rated player doesn’t always get an extra white.

13

True, but that is an inconsistency in Chapter 12 because the last-round coin toss is only for 4-player sections, not 6- or larger. Chapter 12 should have the same rules regardless of number of players.

So:

  • Leave Chapter 12 as is, except remove the coin toss in the last round for quads.
  • Change 30G to be consistent with Chapter 12, i.e. state that player numbers are to be assigned randomly in quads, just as they are in larger round robins.

That way, the need for a last-round coin toss in a quad is eliminated, because the player numbers have already been assigned randomly anyway.

And the organizer can implement randomness in a quad with a single coin toss, as I pointed out above. Heads, number the players (in each quad) in order by rating. Tails, number the players in reverse rating order.

Bill Smythe

14

I think what should be the point of the manual system is that it shouldn’t matter in which order you fill out the lines. If you flip for last round color, there is no effective difference among positions, so you can add people by rating order, or by order they come in the door, or you can switch player 1 from quad 2 with 4 from quad 1 to give different groups than last week, etc.

15

Maybe I am missing something obvious, but toss-for-color in the last round of quads has been near-universal for the last 30 years from what I have seen.

When I started playing rated chess in 1980 there were complaints that the quad pairing tables in the 2nd Edition were unfair and the only way to make them fair was to toss for color in the last round. So that became the de facto standard, which I believe was reflected in the 3rd Edition.

Making the highest-rated player in a quad number one per the pairing chart is also tradition, but I have seen at least one case where that did not happen, perhaps for the reasons Tom M explained. But I cannot remember the last time I played in or heard first-hand about a quad that did not feature toss-for-colors in the last round.

16

If pairing numbers are randomized (i.e. you’re running the tournament by computer), then the colors are automatically randomized.

17

Is WinTD so poorly written that you can’t assign pairing numbers in a round robin? Also, are round robins so complicated that even the meanest club TD can’t pair them by hand? I don’t see the advantage of using software to pair a quad.

Alex Relyea

18

Only tossing in the last round isn’t quite the same as the total randomization techniques afforded by randomizing player order. If you had the same four players every week and their relative rating order didn’t change, they would still play the same colors every week in rounds one and two. Randomizing the order listed or the colors for every round via a single coin toss is much more effective.

19

Yes, you can assign the pairing numbers. Why you would is unclear if the players are in the computer.

If you’re managing 10 quads, there are substantial bookkeeping advantages in having everything computerized. You can, e.g. generate the same types of sheets for players entering results that you would in a Swiss. (One pairing sheet across all sections). A lot easier to get the results in than transcribing from a bunch of separate pieces of paper. (I’ve done the latter on a large set of RR’s—and I would never want to do that again).

20

I agree. But I understand the reasoning behind manual systems.