We had a round robin for our club championship last year. Random lots were chosen for pair numbers and then the Crenshaw-Berger tables were used. In several cases, a player got black against most of the high rated players in the group and white against most of the low rated players. I vaguely remember a discussion about how this was unfair, but that didn’t prevent me from setting up another round robin championship this year with lot numbers. The Senior TD said I set it up wrong and that the Crenshaw-Berger tables are designed for rating order numbers, not lot numbers. Most people got even distribution of colors against top-rating people, but #2 (in rate order) got black against #1 and #3, but he has white against #4 and #5. The rulebook specifically states that round robins are set up by lot number while swisses are set up by rating order numbers. Opinions on the legitimacy of lot numbers in round robins?
Ernest:
Lot numbers are given or just assignment from the director, if the tournament is a round robin and not a USCF rated event. Assignments are more given for pairings for first registered first ranking, second registered second ranking, ect… That is fine if the event is a non-USCF tournament, when this happens in a blitz event or quick events.
Since the event is going to be USCF rated, the round robin crosstables will be designed in the member service area in standard final score with rating only being higher if the final score are the same. Since it is more common for a 2000 player having a higher final score then a 1200 player in a tournament – if the event used the “lot system” a 1200 player could be seeded as number 1 and the 2000 player could be seeded as number 20. Sure, it could be a problem with the higher rated players having a better time with white then with black. Color history is not that much of a problem, it is more the problem of who plays who and when – then the assignment of the color could be a issue.
The “Lot System” for a round robin tournament when it is USCF rated event is ‘Old School’. The lot system was fine for the round robin tournaments during the 1800’s, up to the design of the Swiss system. It left the organizers without the problem of telling the players what their seeding would be, the lot system is just general random assignement without finding who is thought to be stronger or weaker before the tournament.
Since the USCF does have a ranking system with a rating – it has taken that problem away from the organizers and the directors. Just use the USCF rating as the order of the seeding of the players. Like it or not the rating department will design the crosstables with final score, if they have the same final score the seeding will be done with the post rating. The way you had it before, was a lot system converted into final score final rating seeding. The more natural way would be pre-tournament rating seeding into post-tournament final score final rating seeding.
There is nothing wrong with the lot system, only use it for non-USCF rated events; or in the case of having a rated round robin event when nobody has a USCF rating.
The senior TD is wrong. The Crenshaw-Berger tables were intended for use with random lot numbers, as should be obvious from the fact that they were used in international tournaments long before most players had ratings. The color allocation averages out over time.
As a matter of interest, the difference between the Berger and Crenshaw tables is that the order of rounds is reversed. This allows a set of corrections to equalize colors if someone drops out. Whether this is worth the trouble is a good question, but there is no particular downside to the reversal.
This TD is wrong, as your reading of the rulebook is correct.
If the players are paired in rating order, then the top half of the group (the higher rated players) would have one more white than black, in even-fielded events. This was always the reason I found for drawing lots, but now John gives us another reason, which is:
Pairing numbers are assigned by lot at the beginning of the event, unlike Swiss tournaments in which pairing numbers are determined by rating (page 293, Official Rules of Chess).
The ‘lot system’ does work, would even say to John and Terry have more events as round robins rated events then the two combined. Have been thinking of the same problem that Ernest and the players have been asking; even that the color assignment would be random with a lot system or the rating system; random assignment of color assignment would feel unfair for some of the players.
The reason for myself having the round robin, as a number of my events have been quick events with just a small amount of players. Whats unfair for Ernest and the players, is having one game per person. The lot system, when having one game per-person with a number of players. Having 21 players in a event would force the 21 players to play 20 games, making each player have 10 games as white and 10 games as black. Some player could say when they have black, they are faced with players much stronger then themselves – making the color assignment for the player having black or white very sure they would lose the game; when they have white they are faced with the much weaker players, knowing they would have a very good chance to win whatever the color assignment.
The only way to make it fair for Player A and Player B with color, would be having equal amount of white and black between both players. This is the reason for having the ‘match system’ for the players. The ‘lot system’ or the ‘rating system’ still gives random assignment of color; the ‘match system’ gives both players equal amount of color of white and black for both players.
If having 3 people in a round robin, say Forsythe, Player A and Player B. Would pick a even number of games, in this case 10 games. Forsythe vs Player A with 10 games; Forsythe vs Player B with 10 games; Player A vs Player B with 10 games. In this case study, Player A is 200 rating points lower then Forsythe; Player B is 250 rating points lower then Forsythe; Player A is 50 points higher then Player B. If it was a ‘lot system’ with a single game: Player A could have gotten white against Forsythe that is 200 rating points higher; just to get black with Player B that is 50 points lower then Player A during the next game. For Player B would want to get black against Forsythe and white against Player A – as a one game match up with Forsythe vs Player B would be the difference of 250 points making it more clear the game could be won by Forsythe whatever the color assignment. Between Player A and Player B, with only a difference of 50 points – Player B would have better chance with white to win against Player A even if Player A is higher rated by 50 points. The color assignment is more important between Player A and Player B, as their ratings are only the difference of 50 points.
If it is a match, Forsythe vs Player A, Forsythe vs Player B, Player A vs Player B – all the players would have equal amount of whites and blacks. It is much faster for the director or the players, just have one meeting between the players and have them play a set number of games – being equal with each other with the number of whites and blacks. It takes care of the problem of having a single round with a ‘lot system’ or the ‘rating system’. As color assignment or the rating difference would be removed from the problem.
The single game between players in a round robin is unfair for a number of players. Picking the players in seeding with a lot system saves the problem of the director or the organizer of seeding players; when the players can check each others rating between each other knowing what could happen with the out come of the game. The ‘lot system’ works best when the players have little understanding of how stronge they are before the start of the game. When players know each others ratings before the start of the single game between each other, could make both players hate the color assignement each gotten during that single game.
If the round robin event is a closed event, if the organizer used the rating for the seeding – the players would know before the day of the tournament the color allocation with each player they would meet. The ‘lot system’ would be best: it would take away from the players the information of what color allocation they would have with the other players. That would take away the pre-tournament study of what opening a player would use against the others. That is the reason for the ‘lot system’, not the reason for Mister Hongs question.
As Mister Hong having a ‘open round robin event’, the players would not know how strong the other players would be in the round robin. The problem with weak and strong players meeting, if the ‘lot system’ or the rating of the players would still make the color allocation unfair for a single game. If we have a theory round robin event with seven players; the event with seven players would force each player with three games with white and three games with black. The color allocation averages out over time; the rating allocation averages will not work out with color allocation in this theory. In this theory event – it has six masters with one Class D player; six masters will play against a Class D player; three masters will get allocation with white against the Class D player, three masters will get allocation with black against the Class D player. The six masters should have little problems to win against a Class D player with white or with black. In theory the Class D player could win the event against all six masters: in this theory the Class D player lost all the games.
If the three masters that play against the Class D player have white, they will have two whites with the other masters, three blacks with the other masters. If the three masters that play against the Class D players have black, they will have three whites with the other masters, two blacks with the other masters. This is the main reason why round robins have been ‘closed events’. As Ernest Hong had a ‘open round robin event’, it would force stronger players to play against weaker players; if white is eight hundred points stronger then black, white would feel cheated of having white as black is much weaker – as white would feel it should win with any color allocation.
With case study of this tournament, three masters would feel cheated out of having white – as they now would face the other masters in two whites and three blacks; the three masters that got black would now be faced with having three whites and two blacks. In the end the color allocation works out even for all players, three whites and three blacks. The three masters that got black against the Class D player would have beter chance to win the event then the masters that had white against the Class D player. In theory the three masters that had black against the Class D player should have someone in that grouping winning the event; when the three masters that had white against the Class D players sound not have anyone from this grouping winning this event.
What Mister Hong had as a problem, was having a ‘open round robin event’ with one game with each player. The round robin works best if it is a closed event, when it can give equal amount of whites and blacks for all the players. This is the reason why the ‘open round robin events’ have been replaced with the ‘swiss system’.
A round-robin can equalize Whites and Blacks only if it has an odd number of players, with a bye each round. Very few tournaments find this desirable.
The swiss replaced the round-robin in the U.S. because the round-robin cannot accommodate large numbers of players. An attempt to adapt it was the Holland system (used in the U.S. Open before 1950, and in the Olympiads until fairly recently), with preliminary and final round-robin section. But once ratings became common (allowing reasonably accurate seeding), the swiss simply worked better.
Lot numbers are preferable.
In the “average” case, players will receive white or black about an equal number of times against the high-rated players, and ditto against the low-rated. Whenever anything is done by lot, there will always be a few exceptions to the “average”, but that doesn’t affect the legitimacy of the concept. Analogously, a coin toss is not unfair just because, once in a great while, 100 coin tosses will come up with more than 60 heads or more than 60 tails.
One trouble with going by rating is that player 1 always has white against player 2, regardless of the total number of players. If you have two strong players in your area who always come to your tournaments, it is likely that one of them will always end up with white against the other, week after week.
If you wish to remove the randomness (and with it the possibility that a player or two may receive a disproportionately high share of the same color against the strongest opponents), you could try another idea that combines the advantages of both methods. Number the players by rating, but flip a coin ONCE at the start of the event, to determine whether to reverse all the colors (compared to what’s in the C-B tables).
Bill Smythe
John Hillery:
As ratings are common, if you’re in a round robin event, would you feel cheated having white if the other players rating is 1000 points lower then yours; even knowing all the players would have equal amount of white and black in a single game per-person event?
Bill:
That can happen with the top players in the seeding, if having the two higher rated players come to the round robin event(s). Ratable round robin events are very uncommon, the norm would be having round robin events if the event is a non-USCF event: like the weekly blitz event at the local chess club.
If it is a ratable round robin event, with the best and equal way for having equal allocation of white and black for all the players, with equal allocation of white and black between two players – would be a round robin with two sections. That would give the players equal amount of whites and blacks with every other player. The problem would be doubling the amount of games to settle the collor allocation problem.
The only way to make a round robin fair for all the players, with the issue of color allocation would be having a round robin so the players meet each other in a even amount of games, with equal amount of whites and blacks between the two parties in question.
I’m afraid what you’re saying doesn’t make sense to me. You have not explained how how players in a round-robin can have equal numbers of White and Black. If you are comparing a round-robin to a swiss with an even number of rounds – so what? As to whether I would feel aggrieved at “wasting” a White – probably. But this happens in tournaments all the time (swisses too), and I learned a long time ago to get over it and wait for the next tournament.
You are, or course, correct in your later posting that double-RR is the preferable format, but most tournaments don’t have enough time for this (except in things like blitz/quick).
John Hillery:
Say you’re in a round robin to give you one white and one black for both players. Please turn to page 294 of the United States Chess Federations’ Official Rules of Chess, 5th Edition, there on ‘Table A’ is the pairing order for a one game order for three or four players. If the tournament has four players, this will be the order to have one white and one black from the same person.
Round 1 Pairings 1 - 4, 2 - 3
Round 2 Pairings 3 - 1, 4 - 2
Round 3 Pairings 1 - 2, 3 - 4
Round 4 Pairings 4 - 1, 3 - 2
Round 5 Pairings 1 - 3, 2 - 4
Round 6 Pairings 2 - 1, 4 - 3
If you’re seeding number is number one, you will have white with player two in round three with black with player two in round six; you will have white with player three in round five with black with player three in round two; you will have white with player four in round one with black with player four in round four. This will give you one white and one black with player two, player three and player four. This will give you three whites and three blacks with this type of event.
If in the case you only have three players:
Round 1 Pairings 1 - bye, 2 - 3
Round 2 Pairings 2 - bye, 3 - 1
Round 3 Pairings 3 - bye, 1 - 2
Round 4 Pairings 1 - bye, 3 - 2
Round 5 Pairings 2 - bye, 1 - 3
Round 6 Pairings 3 - bye, 2 - 1
If you’re seeding number is number one, you will have white with player two in round three with black with player two in round six; you will have white with player three in round five with black with player three in round two. This will give you one white and one black with player two and player three. This will give you two whites and two blacks with this type of event.
If it is a single event, with equal amount of games:
Round 1 Pairings 3 - bye, 5 - 4, 1 - 2
Round 2 Pairings 2 - bye, 4 - 1, 3 - 5
Round 3 Pairings 5 - bye, 1 - 3, 4 - 2
Round 4 Pairings 4 - bye, 5 - 1, 2 - 3
Round 5 Pairings 1 - bye, 2 - 5, 3 - 4
This is a example of a round robin, the players would have two whites and two blacks. If the event is going to be closed, as most of the round robin events are at the top level of players – this would be an example of having equal amount of whites and blacks.
Of course you can balance the colors with a bye each round. I pointed this out at the start of the thread. Most organizers do not want to do this, since everyone wants to play each round. Likewise, you can balance the colors with a double-RR, but who has time for this in “real” chess? Both of these points seem too obvious to be worth making.
With that, you have given the answer for the decline of round robins. The information of my last post would be fine for a two day or three day event with long time controls. The only organization still in the use of a round robin events, would be FIDE. Having a round robin event, can be done with classical time controls: its’ not that important for the organizers or the players. Organizers are looking for a large turn out, with a set number of rounds. Round robins, are used in the states for just the fun club tournaments of the blitz events.
Can anyone show any event being a round robin, open to the public; being a open round robin without knowing on the day of the event how many people or rounds; being at the classical time controls without knowing the ending date of the event?