Either the game of chess should be a win for White with best play, or only a draw. Assuming that it should be a draw, I would not concede there is any reason for awarding more than a half point or less than a half point for that result.
If we’re talking about God versus God, I’d have to agree with you. But for us finite types, there’s no reason why the scoring shouldn’t take into account the practical advantages of having the first move. To be sure, there’s no reason why it has to either.
Of course, then there’s the practical advantage that Black often has of having fewer lines to calculate…
Having the first move is part of the game and it evens out over the course of an event with an even number of games. I have previously pointed this out in this thread. I do not see the need to tinker with the manner of scoring. I suspect you are taking an abstract notion into account.
Aak! Can I get .7 of a point when I am up a Rp and a wrong color bishop? What if I have bishop + knight vs. knight; can I get at least .65 of a point for the draw? What about all of those silly hundreds of rook and pawn endings that end in draws? Can I get an extra .2 for every extra pawn I have? And Queen endings. How many of those are drawn? Can I get extra points for drawing an ending when I am down a Queen? How about perpetual checks? Should they not be given special consideration? By gum, we need a commisioner of draws to determine how many points a game should be assessed.
Or we could leave things the ways they are. One point for a win. One half point for a draw. Zero points for a loss. Seems like a pretty simple and scientifically elegant way to apportion results.
Sure, for an individual game or a match. For the Swiss tournament, where it kicks the can of complexity down the road into various tie-break systems, maybe not so much.
At any rate, when I started this thread, I was contrasting the .6 /.4 scoring to the 3 points for a win, 1 for a draw method, which Ballard and some others used in a few tournaments. If you’re going to award more for a win and a loss than for two draws, maybe you shouldn’t award so much, and it seemed .6/.4 more closely reflected the actual win/loss statistics.
As I think about it, maybe the idea would serve better as a primary tie-break system. It wouldn’t complicate pairing, and seems fair, since it doesn’t depend on opponents’ subsequent results.
An excellent idea.
I don’t think I want to be around when you try to explain how those tiebreaks were computed to the upset parent of some distraught child.
This, of course, would be another example of a tiebreak system that cannot be computed from the crosstables on MSA, because it may not have the ratings that were used to compute them. (And I’m not sure what to do about unrated players.)
Does it seem fair? I suppose another question to be answered is what tiebreaks should measure?
If you’re talking about a flat score for drawing with Black of .6 - then what you’ve done is ignored the strength of the competition. In one way or another, other tiebreak systems measure this in order to show who played the strongest opposition in order to break a tied score. So why is it “fair” for a tiebreak system to ignore the opponent’s strength/performance?
If you’re talking about a relative measure (one person in the thread suggested scoring based on relative strength - which is really a form of odds) then again it fails for a similar reason. Two players could be tied for a class prize. The two players are rated 1750 and 1650 respectively. Four out of 5 opponents for each are exactly the same, with each achieving the same score against each of those 4 opponents. For the 5th opponent the 1750 plays a player rated 1800, the 1650 plays a player rated 1750 (and since they have tied scores, they have the same results against this one different opponent.)
In this case the 1750 played opposition that in total was higher rated. Yet the 1650 will win on tie-break because the relative rating difference was greater for him(her). Yes, this ignores other scores/tiebreaks - but that’s what the suggestion does - it looks only at rating.
Why does this seem fair?
All in all, this sounds like a bad idea, and one that gets worse the more we think about it.
P.S. To Joe Lux and Tim Sawmiller - sometimes ideas are just bad ideas and need to be evaluated as such. That doesn’t mean we’re belittling the idea or the person.
Unless I’ve misunderstood something, the proposed tiebreak system doesn’t depend on ratings at all, only on results and colors.
In fact, it might work out to be equivalent to the “color” tie-break system, which gives the nod to the player who played the most blacks.
If it doesn’t depend on ratings, though, or on opponent’s playing strength in any sense, then, as Kevin pointed out, it’s probably a bad tie-break system.
Bill Smythe
Bill, I was referring to this (probably bad) idea:
OK, now I see it (way back up there on the previous page).
Probably bad? I’d say definitely horrible. Some people never get over their desire to punish high-rated players for their high ratings. It’s class prizes run amok.
Bill Smythe
Prolly the second half of my most recent post prior to this makes more sense now also.
rating difference points awarded
0-80 0.5 0.5
81-160 0.6 0.4
161-320 0.7 0.3
321-400 0.8 0.2
(Still maintain 1 for a win, 0 for a lose. This should not even need to be stated.)
My system is to encourage the stronger to fight for the win more. It is not to punish higher rated players. (Why would I devise a system to punish myself?) The pairings would be no harder than under any other system. I just want people to play as if they deserved their rating.
The agrument that less draws would make chess more exciting is a fallacy.
Just to be silly: Why do we not learn from football. If the game ends with a draw, restart the game and the first to check or capture wins. This would give the white player a won game.
Um…because football isn’t chess? Because it doesn’t fundamentally make sense?
Why should scoring for wins and losses stay the same?
And yes, the pairings would be MUCH HARDER. If board one draws in round one, how many people will be in the .2 point group? How many in the .5 point group? How many in the .8 point group? Pairings are made by score groups, and your system has the potential to significantly increase the number of score groups in a tournament.
Right now there are at most 3 score groups after round 1. In your system this doubles to SIX score groups after the first round.
It makes no sense.
Not only that - but the players rated close to each other would get .5 (say the “C and B” players) but the top player who drew would have .2! So he’ll be “at the bottom” of the people with a draw. What sense does it make to change his ranking, rather than having him at the top of the draws?
Also - consider this game - at the time Black was 2550 and white was 2100. Black even declined a draw offer late in the game until White forced the draw. Why would Black be punished in this game - he came up with excellent tactics to save the game! cavemanchess.com/Games/Bachl … s/base.htm
The idea is really bad. Sorry.
Indeed. I used to suggest to my opponents that we should agree to two draws. Then we would each get a full point.
Actually, that one made a lot of sense to begin with.
Your example could have been even simpler, though. The 1750 and 1650 could have scored identically against their identically rated opponents in rounds 1-4 (as in your version), but then they could have been paired against each other in round 5. Then a draw would automatically give the 1650 better tiebreaks, because his fifth opponent was rated 100 points higher than his opponent’s fifth opponent. 
Bill Smythe
32,000 computer simulations indicate that if only one pair of players draws, the higher player plays the player seeded one lower than normal. As the number of draws in the tournament increases, the higher rated player plays a higher group of players. The inverse is true for the lower rated player.
32,000 computer simulations of a pairing program indicate no significant difference in the time required to pair opponents between the two systems. Since the most common chess pairing system involves pairing off the perfect scores, then pairing the “odd man out” to the next lower score bracket, there is only two brackets, maybe three if these players played before, in play at any time. Therefore I do not understand how pairings would be more difficult.
My system would address the three problems that the current system has. It still has the rare problem of two or more people entering with the same rating.
Simple example.
16 players with 1-8 paired against 9-16.
Boards 5 and 6 draw, each with a 150-point rating difference.
Standard Swiss has (before color adjustments): 1-4, 2-7, 3-8, 5-14, 6-13, 9-12, 10-15, 11-16).
Your variation where 5 and 6 get 0.4 while 13 and 14 get 0.6 has the following pairing: 1-4, 2-7, 3-8, 13-14, 5-6, 9-12, 10-15, 11-16.
Things get stranger as multiple rounds go by. If you want to do a simulation, check out what happens for a person with two draws versus a win and a loss. At that point you start seeing bigger variations.
- You haven’t demonstrated any issue with the current system
- You haven’t demonstrated that your suggestion deals with any current issue.
- you “32,000 computer simulations” make no sense on multiple levels: first - the results don’t make sense as Wiewel shows. Second, using only one draw doesn’t make sense. Think of the negative impact your suggestion would have in a relatively small 100 player event. Then think about larger events. Third - why would 32,000 computer simulations be needed for only 1 draw? What did you model? 1 draw per round? 1 draw per tournament (and therefore multiple rounds with multiple results and therefore more simulations by changing what round the draw occured in?) As given, the statement makes no sense - if you had a draw in round 1, for example, and you looked at events with 4-100 players, that would denerate 97 simulations. If you went up to 1,000 players that would generate 997 simulations. So where did the other 31,000 simulations come from? Did you rotate the draw among various pairing slots?
4, A counter-example was provided to your system showing 2 players 450 points separate fighting very hard to avoid the draw - yet a draw was a very reasonable result - and the higher reated player who found some stellar moves would be punished by your suggestion - you haven’t indicated why it is reasonable to do so.