My proposed rewrite of Chapter 12 is here and my suggested modification to the original Berger tables is much easier to understand than the current Crenshaw structure.
For a side by side comparison of the Crenshaw and Porter tables, click here. In that model players with consecutive pairing numbers are seated in every other position. By listing the pairings at the odd tables from left to right then the even tables from right to left, players with consecutive pairing numbers will usually be listed next to each other in pairing tables. In my pairing generator rrpair.htm, the option Sequential Players shows how the pairings are currently displayed in the rulebook.
My experience with Round Robins is in my blog starting here.
I respectfully submit this to the Delegates and Rules Committee for their consideration.
In the current rulebook on page 296, the phrase “playing four games:” should read “playing five games:”.
Warren, it will take me some time to study your proposal, and I’m not sure yet whether I’ll support it. I remember playing in a blitz tournament where the pairings were very simple: one player was the anchor and sat in the same place all night and the other players rotated around him. There was no need to consult a pairing sheet or crosstable to know who you were playing each round. If this is similar to what you’re proposing then I agree in principle but I’d like to make sure all the details check out.
By the way, last night I found a mistake in Table F, page 298, in the 5th edition rulebook. This is in round 1 of a 13 or 14 player round robin. The first matchup is shown as “7-4” but it should be “7-14”.
I caught that one a few years ago when I was working on my webpage to generate round robin pairings as well as the one on the next page where in round 1, 15 has white against 1 in the last column. Tim knows about those two already but the error on 296 (“five” instead of “four”) I only found a few days ago.
Bob, I invite you to check out my suite of round robin webpages from the links in my first posting. I am quite familiar with the “anchor and others rotate” idea as this is the basis for round robin pairings. The “By Board Number” option demonstrates this on my webpage. By displaying the odd boards from left to right and then evens from right to left, the pairings in the USCF Rules are replicated (“Sequential Players”). My main pairing page no longer supports the Crenshaw pairings, but a testing page I use to quote pairings in another forum still does: crenshawda.htm.
This is a version of my round robin website with USCF defaults that replicates the tables currently in the Rulebook (with the exception of 4 and 6 player events) uscfrr.htm. This was an earlier thread on the topic from a few years ago.
Either there was a transcription error within the last 110 years or Johann Berger did the short ones in his head before deciding he needed a working algorithm.
It seems to me that the basic problem is that Rule 30 seems to require the use of the Crenshaw-Berger tables. And Chapter 12, which presents the Crenshaw-Berger tables, also states that these tables “are used” for RR tournaments, which seems to imply that this is required, and that there are no other options.
Rather than replacing one set of required tables with a different set of required tables (the Porter-Berger tables), it seems to me that as far as 5th Edition revisions are concerned, Rule 30 should simply be changed to remtove the implication that the Crenshaw-Berger tables are the only ones possible or which may be used. It should be made clear that the tables in Chapter 12 represent one suggested (and traditional) system for handling color assignments in RR tournaments, but that a sophisticated TD with an understanding of the problem may use alternate tables. Somebody who wants to use the Porter-Berger tables for a RR tournament should not have to worry that he is violating Rule 30 or Chapter 12, or feel compelled to announce a “variation”.
For the next edition, perhaps the set of tables printed in the book can be changed. But I would like to see an analysis that shows that the Porter-Berger system is superior to the Crenshaw-Berger system in equalizing colors and is as robust in dealing with withdrawals, late addition of players, etc.
This is another example of TMI (Too much information) in the rulebook. I don’t think the USCF actually cares that much exactly what RR pairing tables are used as long as they have certain required characteristics (such as equalizing colors for the players overall, so far as mathematically possible; maintaining rough color equalization throughout the tournament, so that nobody has a long run of games with one color and so that a withdrawal does not create big color inequities, etc.)
In creating the dropout tables for Porter-Berger, I swapped (10 player example) the pairing numbers 1-5, 2-4, 6-9, & 7-8 with a macro, then sorted the rows of the table since they first came out 5,4,3,2,1,9,8,7,6,10. The tables themselves were created by “seating” the players differently but the rotation around the long table down to the fixed color assignment on each board was identical.
As far as I know no one had done any work on how to add player(s) to a Crenshaw round robin, but I explain how it’s done in Porter-Berger here: late.htm. Unless it’s the trivial case of a last player joining a RR with an odd number players late (becomes the ghost or anchor), the TD had better have a thorough understanding of how the RR works before splicing anyone in after the 2nd round has started.
Of course I would support changes you mentioned in rule 30.