Nolan: Since you are writing the new ratings program, I’m sure you can settle this issue.
It’s the age-old question about losing points by beating much lower rated players. I’ve seen many examples where a player like a scholastic coach rated 1600 beats four students rated 600-700. He might gain 1-2 points.
But in a tourney where a 1600 player has played three rounds – beaten a 1700, drew and 1650 and drew an 1800. For some reason in the next round, he’s paired with a 900 player.
Using the Ratings Calculator, that player may have gained XX amount of points after three rounds. But even if he wins the fourth round, the calculator may deduct a point – almost like he lost a point by winning.
The question - is the guy better off skipping that 4th round and going home? Or is that 1-point deduction just a quirk of the Ratings Calculator.
Well, there is a case with the “special formula” which is for players who have less than 8 previous tournament games. For instance, say that coach you mentioned played in a local quad and earned a provisional rating of 1800 based on 3 games. Then if he playes four of his students in a separate scholastic/club event (make them all rated 600 to make the explanation simpler), and wins all four games, his tournament performance for those four games is 1000. Normally, you cannot lose rating points for going all wins, but in the special formula case, since the coach only had three previous games, his rating will go down even though he won all four games!!! I personally don’t think this makes any sense, but according to the documents on the site, and the ratings calculator, thats the way things are (correct me if I am wrong). Players should not be losing points for winning all of their games no matter how few they have played before.
However, in your situation, I think its just an inaccuracy of the ratings calculator. Once the player has an established rating, there should not be such instances.
If the 1600 player is going into the pairing with round four with a two point score after round three, so would the 900 player if there was a even number of two points after round three.
It is only a factor for the provisional players to lose rating points for a win against a weak player; the fewer the games with a provisional player the wider swing in rating points. If you do have that 900 player with a two point score after round three, then that players rating will be going up much faster then the 1600 player.
If the 1600 player is a established player, would not even worry about losing points for a win against a 900 player. Would be more worried being paired up with someone half my rating with a two point score out of three.
Sure, with the first few games a player is in this could happen. If a unknowed rated player lost all their games against Masters, would have a higher rating after the tournament then a player that wins against 100 - 300 players.
If a player should be at a Class C rating, going to a tournament and play players with ratings at the most bottom end, and they do play at that rating – then the player had less risk and should show their rating with the amount of risk. Try this out, if a international master from a different country is only 12 years old. Comes to America and play in a scholastic event with the top seeded player at 500: the 12 year old international master from this different country should win every game. The player did not have much risk to lose games, the risk factor should show the rating.
If the unknowed player lose to a master, the unknowed player would have a higher rating. The risk to always play against masters is so low it is not even debated, unless that unknowed player is that 12 year old international master from a different country. With having so many games to play with different types of tournaments, the type of person they meet will even out after time.
You are completely missing the whole point. We are not talking about why you are playing the lower rated player, but why a win would cause you to lose rating points. You talk about risk and everything, but say the tournament was a round robin; then you would have to play every other player whether you loke it or not. In that case, we want to know whether the player would or would not lose rating points by beating a low rated opponent.
A player with an established rating should never lose points for winning all of his or her games and should always gain at least one point from winning all of his or her games in an event. (That’s because fractional points are rounded up at the end of the formula, or more correctly, the ceiling of the fractional result is used, so that a gain of .25 is rounded up to a 1 point gain and a loss of .25 is rounded down to a 1 point loss.)
However, the ratings formula takes into account the expected result for all of the games played as a group, not each one individually. Therefore, winning three games might well produce a slightly larger change than what would happen if a fourth win against a much lower rated opponent was thrown into the mix.
The ratings estimator gives a reasonable approximation of someone’s rating, it is not intended to exactly duplicate the ratings formula in every aspect, that’s why it’s called an ESTIMATOR!!
Weightlifter the problem is this. It does not matter what your rating is in the first 25 games, or the first 50 games. After a player has had 150 recorded games, then it will all balance out to a standard range.
WHAT??? Where did you derive this theory? Did you just make it up? Nowhere have I read that one’s rating will balance out after 150 games. I have seen players constantly improving (not leveling at a plateau) even after they have played 500+ games. And did you say that your rating after 25-50 games does not matter? What kind of reasoning is that? Everyone knows that anything after 25 games is considered an established rating, which is used for top 100 rating lists, and it also determines your section in large class tournaments. How can you say such a thing as “It does not matter what your rating is…” No matter what your opinion about the rating system is, the reality is that it does matter.
Just because a certain scenario applies to you and maybe a few other players you are familiar with does not mean that it is universally true. If you have any reason to believe that what you say has any evidence to back it up, please show me.
None of what you have said has answered CRW’s questions. If you have new ideas or suggestions, that is great. However, make sure that what you say is completely valid and that you can back it up with proof. To make an observation is one thing, but to jump to incorrect conclusions is a different story. Most importantly, make sure that your ideas are relevant to the issue at hand.
True, it is an estimate, but on the website, it’s actually called the RATINGS CALCULATOR, which could be misleading. Maybe it should be changed to RATINGS ESTIMATOR like you say in order to clarify that its calculations are not so accurate after all.
I believe that the following statement in the ratings calculator program is sufficient as a disclaimer.
We get a lot of questions about ratings changes, most of them from people who think their rating should have gone up more than it did.
I have yet to come across an example for an established player where the ratings calculator was off by more than a point once it was given accurate data, and I think those are probably roundoff differences due to the fact that the two programs are written in different languages and thus use different math libraries.
GM Andrew Soltis came out with a great statement. It was a number of years ago, his statement when a player reach the best they will ever become would be seven years as a tournament player.
Lets look at it this way, if you come only to my tournaments (2005 calender year) as being a four round and one tournament a month, would have only fifty-two rated games in a given year. That would be very active for any given player, most player at best looking for twenty given games in a given year and that would be very active. When after having one hundred and fifty rated games, the player would have settled down and find their norm with their personal rating. Only talking about rated USCF games not the games you play with a friend or a computer. They are nice to play with a friend but does not change the formating that only rated games are debated. Can show you a number of players that over the last seven years, have not even been close to having one hundred and fifty rated games. For these players, if they can go to any tournaments and play one hundred games in a given year: there rating will not go up or down in the wild swings of a new player.
For the bulk of players, the rating does not matter if it does not add up to a title. The USCF rating is not the important rating, its’ the FIDE rating that is important. Needing a good FIDE rating just to get the norms for a FIDE title, needing a good performance tournament just to earn these norms. Sure, USCF does give titles with ratings; will give respect to a person with a higher rating then myself, like a expert or master or someone in a higher rating class then myself. They only give respect for the persons class or title of expert, master, senior master.
Even if a person wins a FIDE title, like FM, IM, GM, WFM, WIM, WGM: these titles are only good and useful if they are used in the society of chess players. Weightlifter, when you get a title even titles earn from college your in a pidgen hole with what you got and have to deal with it. Try it this way, look at GM Larry Evans. He has been doing a chess colum for generations, talking about old times and what Fischer ever did. Have let to read the January 2005 Chess Life, but can say he will have something to say about Fisher. Look, he has a grand masters title, he has the respect of players for his past performance; the factor is this, he is still in a pidgen hole as he can only make a living talking about chess and Fischer. He can write about about football or try to be on ESPN during half time: do not see a super bowl ring on GM Evans finger or know what football team he was on.
Why the rating does not matter, as it does not matter outside the chess society. If going to a grocery store to seek employment as a stocker, the managment does not care if myself being a Class C player or being a tournament director. Do know a master as a taxi cab driver, the people that get in his taxi do care if he plays chess or being a master. They care to get from point A to point B and that is it. If you both were at a tournament and being paired up with each other, then you would be given him respect for being a master.
Fisher? Evens? Do you mean Fischer and Evans? If you are going to say disrespectful things about other people, at least have enough courtesy to spell their names correctly. Do you have a problem with Mr. Evans? As far as I know, he was a great chess player and a nice man. Please, do not say bad things about people that you do not personally know. That really offends me (and many others too I am sure). If you say that GM Larry Evans is stuck in a pidgeon (not pidgen) hole because all he knows is chess, then stop for a moment and think. What about yourself? Don’t you play a lot of chess? Are you not a tournament director? How can you say that chess is a useless waste of time? If you really think so, then how much time have you wasted? Do you have a super bowl ring on your finger? I think not. Yet you do seem to have the time to post over 500 messages on this forum. I sure would rather be in his pidgeon hole than yours.
Once again, please think before you type. Make sure you really know what you are talking about. And most importantly, try not to disrespect others, unless they have personally caused you harm. If Mr. Evans has ever done someting bad to you, go report it to the police.
My youthful young man, as you can see, my youth has past me many moons ago. If you’re fantasy is to be the next GM Evans, will not stop you and you’re dream. Myself only look at chess as a way to keep my mind on something – as life is more then a 40 hour work week. Could be in bed half the day and the other half drinking. Well the work week is 40 hours and there are 168 hours in a week.
Now how did this talk of a few games of chess became a subject of chess being a useless waste of time? There were chess tournaments even before Dr. Elo was born; players did not care of titles before Zar Nickolas II gave out the first titles of Grand Master. Dr Paul Morphy did not worry over what his personal rating was at any given time in his life, as there was no rating system. Dr Paul Morphy did not worry over titles like Grand Master, as there was no title like that at his death. He went to Europe for two years, defeating player after player: then went home and became a doctor.
Chess players up to the time of Botvinnik, the ones we still print and study before Botvinnik: were the upper social class during their time. If you’re looking at chess as the meal ticket, its only and only for a very few players. When you get a house payment, a car payment, house insurance and car insurance; when you get a number of bills you always have to pay month after month. Then you will find going to a weekend tournament is a way to escape from the grinding forces of the real world.