the initial post referred to using handicaps. If you're not using handicaps, it's not like existing point/ranking systems aren't working, so why change it? Lin Dan has like 7000pts last time I checked the standings, if you have 100pts, I can also estimate your probability of winning. One huge downfall of ELO is it doesn't account for inactivity. A player who achieves a high elo rating can simply stop playing and still maintain their status. Whereas current bwf and even tennis scoring systems will punish players who get lazy. At the end of the day, I like keeping things simple, and not trying to fix something that already works... and works well.
I would say for these reasons that the current BWF and tennis systems don't work well. Sometimes we see good players who've missed a couple of tournaments due to injury (which is not at all the same as getting lazy), their rating slips, and so you get a defending champion going into a tournament as an unseeded player. It shouldn't happen this way. Also, if you look at rankings outside the top 100, there's very little correlation between ranking and a player's strength. If someone gets a lucky first round draw in a minor tournament, they might pick up enough points in one match to move them from number 200 to number 150; another equally strong player might get a tough first round match, lose, and get no rating points.
To urameatball (post # 19): As the whole point of this thread was for a single club internal tournament, why shouldn't everyone have a chance of getting to the final, especially if they're playing significantly above their usual standard? However, if you want to ensure the best players will probably get to the final, you can scale down the handicaps. An annual tournament at my old club was billed as a 75% handicap. It then became a 50% handicap when I won it with a handicap of +3. To Gollum (post # 20): The system I'm proposing is to apply Elo to the proportion of points won rather than the overall result. So losing 19-21 to Lin Dan would be a substantial boost to your rating. However, your point about players who wouldn't usually play one another is valid. See posts #16 to #18 about 'separate communities'. To urameatball (post # 21): Inactivity is another example of 'separate communities' (again, see posts #16 to #18). Also, a high rating is a disadvantage. The higher your rating, the more points you have to win to maintain the rating. A player coming back from injury will tend to have too high a rating. If it isn't artifically corrected, they will have an uncompetitive handicap until results bring it in line with performance. I also agree with alexh. Round-based point allocations help top players keep their rankings. For example, the top 16 won't face anyone else from the top 16 until the last 16. Therefore, the 16th ranked player/pair will have more chance of getting to a latter round than the 17th ranked (who may have to play the world champions). They'll then pocket more points, keeping them in the top 16. The best example of this is snooker. The top 16 automatically get entered into all the major tournaments, where as 17 and below have to play up to 4 qualifying rounds. This means 17 and below may only feature in half to a quarter of tournaments where they'll probably play the world champion at some point. It's entirely possible that half a dozen players should be in the top 16 but aren't by the inertia of the system. Elo forces players to perform to their relative strengths & corrects when they don't.
call me old school, but hard work, lots of practice, and playing your best = win. using elo, you're inclined to play poorly until a tournament in order to give yourself the best handicap. I went from having the lowest ranking in a tournament (had to play 1st seed in first game) to becoming a seeded player myself. It was immensely rewarding to see all my hard work has paid off. Under the elo system, I'll feel as if I'm punished for working hard and improving, while slackers are rewarded. tournament finals are so the best of the best can put on a show... and not for the trickster who managed to artificially lower their elo rating.
There seems to be a bit of confusion here. We're combining two different ideas: (1) using the Elo rating system; (2) running a tournament with handicaps. These are two separate issues. It's possible to use Elo without having handicaps. And any sort of handicapping system is open to abuse. Handicaps are not part of the standard Elo system.
In the case of standard tournaments, I agree with urameatball. The best should be seeded, so that they populate the latter stages. I think this is especially true at elite levels. However, I believe that there is also a place for handicaps. They give the 90% of a club who don't have the opportunity to play >5 times a week (kids, job, torn knee cartilage etc) a shot at winning. It's also a great way to highlight the most improved player. A further advantage is that the games tend to be closer & more of a spectacle. Those who would try to artificially lower their rating can be an issue. However, this can be combatted in a number of ways. Firstly, tournament matches could have a higher weighting ('K' factor) that ordinary club nights. Secondly, seeding would mean that anyone who artifically drops their rating could find themselves in a group/draw of death. Thirdly, the handicap could be partial, ensuring that the best will still probably do better. Finally, any sudden dip in rating would be difficult to hide from a half-awake committee. Improving your Elo in certainly less tangible that a trophy. However, you win a trophy because you played well on one day (or in my case, my opponents had a domestic arguement in the semi finals). An improved Elo is the product of consistently playing better over a period of time, which could be argued to be just as important. I am not saying urameatball is wrong. There is no substitute for hard work and long hours. However, a flaw in badminton is that a small difference in ability leads to a fairly large difference in result & genuine contests are rare. Narrowing the gap (albeit artificially) make the game far more inclusive, especially for those who's best days are behind them. I think that there is justification for standard and for handicap. I also believe that there is room for both, so why limit ourselves to one or the other?
I had jumped in, without first looking, to indulge in my favourite pastime. That is, thrash talking. However, after reading some of the posts with interest, and considering the imperfectness of the current BWF ranking system, I feel that much more scrutiny is warranted. Let me use the following terms for the convenience of discussion: Match Score - that is, W or L, Game Score - 2-0, Point Score - for example 21-15 21-10, giving a total of 42-25, and lastly Ratings Score e.g. 2600 I had been slow to read and quick to speak, and being unable to find the earlier posts related to this thread, I may have missed the gist of the discussion. Are you suggesting trying the Elo system just for internal club tournaments? I have been thinking of its possible use in the world rankings of the players. The Elo ratings system was developed primarily for the use in chess, since Elo himself was a chess player. In chess, there is no degree as to how much you win or lose. That is, there is only a match score, but no game or point scores. A smaller ratings score is awarded to the winner or subtracted from the loser if a match result is closer to the predicted score based on the relative ratings, and, a larger ratings score if the match results is further away from the predicted score. This has been illustrated by Line & Length above. Since the predicted score does not affect the match score, but only alters the ratings scores, then it should not be considered a handicap system. The Elo ratings system has, however, been applied in other sports with game or point scores. But I have not yet looked at how it works. What I think Line & Length is proposing below should be a handicap system, since the predicted score will affect the outcome of the match. I'm not sure as I am too lazy to read all that technical stuff. I'm skeptical as to the fairness of this system, or whether the stronger players will be happy with it. What may be possible, however, is to leave the match result as it is and instead use the deviation from the predicted point scores to transfer a less weighted ratings score from the loser to the winner. Currently, the Elo ratings system uses the match score to transfer a much more weighted ratings score from the loser to the winner. However, the Elo system based on the match scores seems to be working, and I believe only a very able mathematician will be able to predict whether the system using point scores can be successful. I think there are two main reasons for a ranking system in badminton. Firstly, to prevent the top players from eliminating each other in the early rounds. And secondly, to qualify for world or continental events, as well as the grand finals at the end of each year. Inactivity can be dealt with easily. The BWF can simply stipulate the minimum number of tournaments (that awards a certain maximum possible points) that a player plays one year prior to the tournament in which he or she wants to participate. Failing which, penalty points may be deducted from his rating score. This will affect the seedings as well as qualification for the major tournaments. Conversely, the current BWF system is rewarding overactivity. Some national associations or companies may be able to send their players abroad for tournaments much more frequently than others. Even a less deserved player may be the world's highest ranked if he or she plays that many more tournaments. For the current system, the BWF can require players to participate in a minimum number of tournaments as mentioned above, or may impose a time-dependant depreciation of the ranking points, something similar to blinding out in poker, to discourage inactivity. For players playing many more tournaments than others, the BWF can perhaps multiply the ranking points gained in a tournament with a factor, a fraction that decreases with the number of tournaments played. So as to dinimish the returns for effort, to level the playing field. One of the hallmarks of the Elo ratings system is the increased accuracy of the rating score with increased participation. This is the law of large numbers. The rating score will simply gravitate to a value which reflects the player's true strength. I still believe that the main disadvantage of the Elo system is the inaccuracy of the relative strengths of the players, though perhaps not so significant. Even if the total population exists as a single community, an injection of a large number of weaker players into the population will require some time for their rating scores to be assimilated into the population to form a new bell curve, time that may exceed the playing life of most players. Sometimes, I'm not even sure if I know what I'm talking about .
