| |
BGonline.org Forums
A tale of two elo's
Posted By: Daniel Murphy In Response To: A tale of two elo's (Ken Larsen)
Date: Saturday, 16 June 2012, at 4:54 p.m.
1. In real life tournament backgammon, people don't have multiple ID's, don't arrange and fix matches to manipulate their ratings1, and don't have unethically achieved super high ratings (6008!?)2. So, I'm not seeing how your experience with the rating list on Pogo is relevant to tournament backgammon.
2. As rew points out, a well-functioning ELO system, like DBgF's, does a pretty job of ranking players by PR, given whatever correlation there is between PR and tournament performance (probably a pretty good one), and given the enormous (compared with chess, for example) fluctuations in rating due to variance in results in typical match play.
3. An ELO system also gives more weight to a player's most recent results. Which is good, isn't it?
Compare, for example, two players playing against average players: one wins 10 in a row and then only 10 of her next 20. The other wins 10 of 20 and then wins 10 in a row. They have identical won-loss records. Which player should be rated higher, do you think? Which player will be rated higher?
4. Backgammon rating systems (and others3) can be set up to delist inactive players. DBgF does not display the ratings of ex-members. NBgF and the BIBA current list delist current members without a minimum of recent activity.
5. Systems for evaluating performance aren't exclusive. A Master Points-type system doesn't exclude a rating system, and vice versa.
6. Systems used in other sports may or may not be appropriate for backgammon. People point to tennis. Here's part of wikipedia's summary of the ATP system:
A player's ATP Ranking is based on the total points he accrued in the following 19 tournaments (18 if he did not qualify for the Barclays ATP World Tour Finals):
- The four Grand Slam tournaments
- The eight mandatory ATP World Tour Masters 1000 tournaments.
- The previous Barclays ATP World Tour Finals count until the Monday following the final regular-season ATP event of the following year.
- The best six results from all ATP World Tour 500, ATP World Tour 250, ATP Challenger Tour, Futures Series, Olympics and Davis cup tournaments played in the calendar year.
Once entered, all tournaments, except for the Barclays ATP World Tour Finals, remain in the system for 52 consecutive weeks."Further, some events count more than others. For example, winning grand slam tournament earns 16 times as many points as winning a "Challenger 150,000 +H" level tournament, whatever that is.
The differences between tennis are backgammon are many. There's no backgammon "grand slam." There are no mandatory tournaments. There is no class of professional backgammon players. With a very few exceptions, most players, including highly esteemed world class players, attend very few tournaments. Likewise there are few tournaments attended by large and overlapping numbers of world class players.
Chess, on the other hand, unlike tennis, seems to do well with an ELO rating system, without a tennis-like, ABT-like, Master Points-like performance rating.
Table tennis also, I believe, uses an ELO system, not an APT-like system.
Bridge use a master point system. But the points aren't deprecated with time. No one every loses points. Poor recent performance is never penalized.
Which brings up another point. What is the purpose of a proposed ranking system? In bridge, for example, specifically ACBL, it seems to me rather obvious that the purpose is to give some encouragement to the vast mass of average and above-average players to continue participating in sanctioned tournaments. Proposals for backgammon discussed here, however, seem focused on how to recognize top achievers.
7. In tennis, fwhich has well-developed and well-funded international and national organizations, someone, with authority, and good repute, and money, has gone to a great deal of trouble to create and maintain a ranking system. Elsewhere, Phil Simborg asks why there is still no international backgammon organization. It seems obvious to me that a reputable international organization of the type that many other sports and games enjoy can only grow out of the cooperation of existing national organizations, and needs funding, and someone to do the work. There are few national backgammon organizations, none of which are well-funded with the some-time exception of DBgF, and all rely on volunteers to do what they do. Make the case to them: what does a proposal for an international organization mean to them, in terms of work required, money required, and the benefit to backgammon in their respective nations?
8. Seems to me that a very minor step towards international cooperation would be, as I've suggested and wished for in the past, a combined rating system, starting with, perhaps, the Nordic countries. But even this minor step entails work and cost, and although there would be benefits, it's not clear even to me that they are compelling. It's fine to discuss various schemes, but in the back of one's mind, at least, should be a thought to what the proposals entail, who would do the work, who would pay for the work, and who would benefit.
1. Years ago, DBgF maintained two ratings lists, one for tournament competition and one for "all matches." The all matches list, among others, included matches that were privately arranged one-on-one encounters. So the all matches list was surely vulnerable to abuse in this way.
2Currently used backgammon systems don't eliminate the possibility of self-inflated ratings. One could, for example, choose to play only in tournaments where, because of the ratings of those attending, one expected one's own rating to rise, with expected results. For instance, a middling PR player who has been unusually lucky in recent competition, who wished to protect his unusually high current rating, might be wise to avoid entering the next Masters event.
3Here's how FIDE handles inactive players:
7.23 Inactive players are not included on the list but nonetheless are considered rated at their most recent published rating for rating and title result purposes.7.23a A player is considered to commence inactivity if he plays no rated games in a one year period.7.23b Inactive players are shown on the next two years of rating lists after starting being considered inactive. Their names are then flagged as inactive in the alphabetical section of the rating list and removed from the national federation’s lists which contain only the list of active players.7.23c A player regains his activity if he plays at least one rated game in a period and he is then listed on the next list.7.24 For the purposes of the FIDE rating list ranking of top players, a player who is inactive over a 12 month period of inactivity on the rating list will no longer appear on the top list.
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.