One issue is the topic of this forum: change the draw score to help prevent tied scores.
My suggestion deals with this issue just as the 0.6-0.4 debate does.
The computer simulation is running a mock 9-round tournament with 100 players. The computer uses percentages to determine what each individual score would be, and is allowed to randomly change up to three scores in the first round. Each subsequent round is allowed more changes, three times the previous round, due to the closer pairings. This simulation was run 32,000 times to allow for a better set of stats. The output was the starting seed of the altered players, the seed number of their opponents under the old system, and the seed number of their opponents under my system. I gave the results of only one draw because that was the highest difference when the higher rated player was paired to a player lower rated than would have happened under the standard system.
Since you wrongly call this a punishment, the 400 point higher rated player should win 99% of the time. Therefore the punishment is for not winning the game.
These debates are getting too far from the forum topic. I just wanted to provide an alternative to the 0.6-0.4 split. A discussion of pairings should be done in the TD’s forum.
That is the issue claimed by the thread. However, it hasn’t been shown that draws are in fact an issue - many of us believe that they are legitimate results and not an issue.
Second, it hasn’t been shown that the scoring change you’ve suggested would change this issue in any positive way. In fact, I’ve provided a real life counter-example showing how it would fail.
And you have exactly the same problems.
What you’ve provided doesn’t give quite enough detail - but it sounds like what you’ve done is diluted the impact of the change by (in essence) comparing it as a percentage of all the unchanged results.
You previously wrote:
What this sounds like is that you took an average across the ranking of opponents. This of course is skewed because there are relatively few changes per person.
It’s like saying: it’s ok if occasionally a flight from New York to Los Angeles ends up in Seattle, because if you look at the average passenger arrivals over all flights for all passengers, it represents very little change from the correct model.
Additionally, it doesn’t sound like your pairing model is correct. What pairing software were you able to use that allowed you to run 32,000 simulations? And if you wrote your own software then there seems to be an even greater issue because it doesn’t sound like you understand pairings as given by this incorrect statement for your system: “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.”
You wrote previously:
This creates NINE score groups, not “3” (I think I incorrectly stated 6 previously because I counted only one side of each score above - my bad.) Do you think you’ll have only one draw in total from each ratings group per round in a one hundred player event?
So yes, pairings become significantly more complex - especially so when players mix their “.6” with a “.3” creating yet another score group.
With win, loss and .5 - there are 3 score groups: 1, 0, .5
In theory a .8 should be paired against another .8. a .7 against another .7. That will happen if they exist. If they don’t exist, they would be paired against a player they are “next to”. The .8 may face a 1, or a .6.
Now think about this next round, if players continue to draw. You could have this many score groups:
Since you’re trying to change behavior by offering a disincentive, that is a correct - not an incorrect - use of the word “punishment”. Players with a rating advantage already expect to win against a lower opponent and view a draw negatively - you’re simply complicating the issue by increasing the amount of disincentive. Do you think in the example I provided that GM Soltis WANTED a draw? He declined one still hoping to outplay is lower rated opponent. But the position is the position - willpower doesn’t necessarily change an objectively correct result.
Pairings are important to understand whether the proposal makes sense. The proposal doesn’t philosophically improve any issue, and is pragmatically a non-starter. The .6, .4 idea is bad, and the proposal you make seems to extend that concept to an unworkable level.
Delegate motion. Why? It appears that no poster is giving an inch in this alleged discussion (it looks more like a debate to me with both sides digging in).
Rules changes affecting pairings or prize distributions that come from people who are not very active players and are not tournament directors or organizers are usually quickly defeated.
Moreover, doing ‘32,000 simulations’ is not the same as running even ONE tournament with players and/or their parents who can complain about either the pairings or prize distributions that would result from the proposed change.
Such as the final round of a five-round event when the sole 4-0 is playing the sole 3.7-0.3 (where the rating difference would mean a 0.6/0.4 split for a draw) and nobody else has more than 3.3?
The scoregroups could be 4.0, 3.8, 3.7, 3.6, 3.5, 3.4, 3.3, 3.2, 3.1, 3.0, 2.9, 2.8, 2.7, 2.6, 2.5, 2.4, 2.3, 2.2, 2.1, 2.0, 1.9, 1.8, 1.7, 1.6, 1.5, 1.4, 1.3, 1.2, 1.1, 1.0, 0.9, 0.8, 0.7, 0.6, 0.5, 0.4, 0.3, 0.2, 0.0 (39 scoregroups).
Or, to go back to the original black/white concept, playing white in the final round with 4-0 versus the sole 3.6-0.4 with no 3.4-0.6, where the .6/.4 split still makes a draw financially feasible.
The scoregroups there could be 4.0, 3.6, 3.4, 3.2, 3.0, 2.8, 2.6, 2.4, 2.2, 2.0, 1.8, 1.6, 1.4, 1.2, 1.0, 0.8, 0.6, 0.4, 0.0 (19 scoregroups).
P.S. the large number of scoregroups makes for smaller numbers of player in each scoregroups, and correspondingly more problematic color allocations (expect to see more triple white or triple black sequences - for the original black/white variance look at four draws in the first round, the four players that had black drawing each other in the second round, and the only two 1.2-0.8 scores paired in the third round with one of them getting three blacks in a row).
Isn’t this thread really about how other people should prioritize and act differently than they customarily do?
When I was young and used to peruse the World Almanac, I noticed that various sports competitions had numeical results that were carried out to several significant digits. And chess, to its credit, gets along with results ending in .0 or .5.
I think the NFL and NCAA football have gone overboard in their aversion to ties. But if we can break ties in a way that gives meaning to the drawn game itself as one of three valid results, I’m all for that.
I do note that the 16-point system I proposed in the other thread (which I really like!) seems to elicit zero interest.
