The Swiss Gambit and Neal Bellon's "The Case for Random Pairings"

I wanted to bring everyone's attention to Bellon's article in this month's "Chess Life" magazine, the topic being his case for random pairings over the Swiss pairing method.  The article is good, not great, but I think it is an important topic that needs to be discussed by real chess players (FIDE, USCF, etc.).

I just mentioned the sort of complaint presented in the article to a friend, before I knew the article was going to be published (it just arrived in the mail yesterday): how often do you ever get to play players rated anywhere near your own rating in a tournament, especially one of those weekend madness tournaments where there might only be one division?  The article's focus is on the major point that the first round is, in most cases, basically a waste of time for everyone.  Strong players play some forsaken lamb or other.  The author (peak USCF of 1779) reflects on a time when he was one of those lambs, put in a situation where he played IM Jay Bonin.  The reason I say it is a good article, but not a great one, is, in part, due to the fact that I don't think a 1700 playing a 2400 is such a bad thing, especially if you want to play strong players.  

I'll take a different tack in the complaint: a real injustice is having to be consistently put up against kids who are many hundreds of points stronger than their rating (because they are coached by top players, learn more quickly than adults, and aren't very active tournament players), and be put in a situation where you are the B, C, or D class player who gets no reward for winning or brutalized (in the points department) for losing and drawing to them.  Jay Bonin rarely has to worry about the B-Class or A-Class player he's playing being significantly stronger than their rating --or stronger than he is!  The random pairings make for a nice change of pace, so that one particular player always paired with such a precocious scholastic player on a regular basis, and it makes it so that a kid rated 800 can steal an early draw with an A-Class, Expert, or Master, rather than consistently taking points from B-Class, C-Class, and D-Class players.  As someone who began playing chess as an adult, this has been a major item of frustration in my tournament career.  Even a few weeks ago, I played a 1200-level scholastic player with just this situation.  His previous supplement said that he was either in the 800's or 900's, so I was completely panicked after he got an advantage in the opening (I know no opening theory) and outplayed me in the middlegame.  Fortunately, the kid recently became an active tournament player and made the jump to 1200 official, so that holding a draw with him didn't cost me more points.  (Sidenote: I subsequently defeated him a week or two after that, once I knew I was playing a 1600-ish player!)  The point is that, in addition to the point Bellon makes in his article, there is probably more of an injustice down at the other end of the rating scale that should force us to critically evaluate using random pairings.

Another problem, not at the Pittsburgh Chess Club or the Old Morgantown Chess Club, but at the Boylston Chess Club, is the "Swiss Gambit" that arises from players: a) not wanting to waste their time playing someone 600+ pts higher, and b) not wanting to waste time AND points on playing someone 600+ points lower, especially when the player is actually stronger.  From personal experience, I see that a heft half-dozen players in the A-Class to C-Class don't bother playing in the first round, which means most of the middle of the wallchart is cut out for the first round, adding disparity to the first round games.  This is the Swiss Gambit, and it generally works out pretty well for players rated 1500-1800, because they get a half-point by, and either play the strongest of the losers, one of the first round draws, another swiss gambiteer, or the weakest of the winners...but usually one of the middle two happens to be the case.  Being one of the mid-level players who wants to play as many games as possible, it really disappoints me to see anywhere between an eighth and a third of a tournament's players not showing until the second round.  What's more, most of the consistent A-, B-, C-, and D-Class players who play in the weekly, one-game-per-week tournaments do not play in weekend tournaments precisely for this reason.  A payer rated 1675 said to me, "Why play a game against one Expert or Master --lose that--, a second against a totally underrated scholastic player you'll never see again  --lose or draw that, or get nothing for winning it--, and then get a game against another scholastic which, even if you win, who cares; your so psychologically beaten before the tournament is over that you withdraw from the fourth round!  And then, of course, the time controls are shorter, which favors kids, who are blitz maniac compared to adults."  Not only would random pairings draw in those players who normally do not play at all, but it would remove the value of the Swiss Gambit.

Bellon talks about a number of other issues, so I'll leave it to you to read the article; but I think he makes many good points about the other end of the rating spectrum with regard to the Swiss.

Again, very good article, not great: I think one of my problems with it is that I don't like absolutist mentalities when it comes to fixes.  I think the fix is not just random pairings.  He admits this, himself, in the article, because he rightly regards the Swiss System as currently ideal for big national tournaments.  What I'd like to see is more variation in the pairing method, in general.  I'd like to see more round robins, random pairings, some ladder-like pairings, or some tournaments that do the first two rounds as random and the remainder as a Swiss setup.  I'd like to see Swiss pairings remain in effect, but I don't want to see them remain the exclusive pairing format.

I haven't read the article but if more players demand different options and voice that concern to TDs and organizers, then it might be something that might change.

I wonder if something like accelerated pairings in the first couple of rounds might solve some of the issues mentioned though.

I was also going to suggest accelerated swiss (aka Dinamo) pairings as an alternative approach.

Ah, how do accelerate pairings work?  I imagine it does something pair players with others closer to their ratings, but what is the protocol; how is that achieved?  A program that chops the field into temporary pieces and pairs that way seems like one way of doing it.

It's like swiss except you divide into fourths instead of in half.  You treat the top 2 groups like it's own swiss and the bottom 2 groups like it's own swiss.

for example

1

2

3

4

5

6

7

8

swiss would be 1v5, 2v6 but dinamo (accel swiss) would be 1v3 2v4 5v7 6v8

Yes.  I've played in this format before and I liked it.  

Also, it's not just class players that have to worry about underrated kids.  I've heard titled players complain about having their rating points drained by underrated 2100-2200 kids.

Personally, I don't care about the whole kids thing.  Some people let themselves get psyched out by it.  I do just as well (or poorly) against kids as I do adults.

A lot of the kids in the 1600-1800 range are overrated in the midwest at least.

PossibleOatmeal provided a good example. It can be used to make sure the higher rated players play each other ealier in the event and is sometimes combined with larger tourneys where that might not happen at all or not until the last round.

I have never actually used them but have seen other events that advertise that they might do it.

"A payer rated 1675 said to me, "Why play a game against one Expert or Master --lose that--, a second against a totally underrated scholastic player you'll never see again  --lose or draw that, or get nothing for winning it--, and then get a game against another scholastic which, even if you win, who cares; your so psychologically beaten before the tournament is over that you withdraw from the fourth round!  And then, of course, the time controls are shorter, which favors kids, who are blitz maniac compared to adults."

I empathize with this 1675 player, but this is very defeatist thinking. It is certainly possible for a B player to draw or beat an expert or master. I have been playing OTB chess for 40 years and know the frustration of losing to an underrated youngster. But it comes with the territory. I think part of the real solution is to become as mentally tough as possible and learn how to play as well as possible against strong players. Learn to love the challenge and don't worry about the pairings!

Thanks.  Yeah, I like that.  It sounds good!

It depends where you live.  If you played at the Boylston Chess Club, where GMs (Dzindzdi, Christiansen, etc.), IMs (e.g., Esserman), and NMs (abound) are all over the place and coaching these kids left and right, you'd probably be complaining.  You really have to hover over the kids to figure out what they don't know, so you can use it against them and beat them.

Actually, this is another problem: the Scholastic players can ruin performances, as happened a couple of weeks ago at a tournament I visited.  A B-Class player defeated a sturdy Expert, then lost to a Scholastic, making the who tournamen mediocre rather than a solid performance.  What to say?  The way the USCF ratings adjustment works, it looks like the fellow lost to and Expert, beat two high D-Players, and an 800-level player --but I guess complaints about the rating system is another story.  

But is it a pairing system problem or a rating problem ? It looks like scholastic players are severely under-rated in the US.

FIDE adressed the problem with the accelerated coeff 40 for youngsters. Maybe something similar is required in the US ?

Definitely, my qualm is more of a rating problem that the pairing system could be a band aid for.  I'm impressed that FIDE proved to be not impotent on the issue with Scholastics.  The difficulty of not getting to play many players in one's own rating can still be fixed by other pairing methods.  And I'm not for trading one absolute pairing format for another.  After all, I score better against Experts than I do against A-Class players.  

Laurent, I realize you probably don't get "Chess Life" in France, so would you like a snippet (within the bounds of copyright law, of course) of the article scanned for you?  Bellon's article addresses the pairing of, for example, 1700's with 2300's, as is typical of many first rounds at clubs like the Marshall or Boylston.

Thanks, but I think you gave a good explanation of the topic at hand : I'm not sure I will find the motivation to dig deeper

If you are advocating for tournies in which you only play players closer to your own rating you are advocating for class tournaments . They already have tournaments that are run like that . The accelerated pairing system is normally only used when there are too many entrants and the number of rounds could produce more than one perfect score . A 5 round event is ideal for 32 players , you cant have more than one 5-0 with a field of 32 in 5 rounds , so if there are say 50 players instead they will usually use accelerated pairings and this greatly reduces the chances of more than 1 perfect score but does not eliminate it completely . I have played for 43 years this year in rated tournies and get the feeling now that USCF doesnt care for adult players and only cares about the kids ... it seems to me that adult players are being gradually forced out of the game . 

That's where the money is.  I don't think it is consciously geared toward eliminating adults.  Despite the buck I spend on chess, the parents of the three kids (from three families) I used to travel to tournaments with would spend a fortune.  When the mom and pop of a household, both well-paid professionals, of chess-playing youths are willing to spend approaching 10% or more annual income on chess coaches, chess camps, etc., the focus of a corporation with interests in capital enterprise will be on those people first.