PDA

View Full Version : stop euchring the adults.



Davidflude
02-03-2008, 10:55 AM
As any adult can tell you juniors tend to be seriously underrated.

If you play events with lots of juniors your rating tends to fall.

Some adults have even stopped playing in clubs with heaps of juniors and switched to clubs with few juniors to avoid this problem.


Even worse the improving juniors will tend to cause average ratings to fall over a period of time.

I have a radical suggestion. That where a junior plays an adult the juniors rating should be altered according to the result but the adult's rating should remain unchanged.

Rincewind
02-03-2008, 11:17 AM
I have a radical suggestion. That where a junior plays an adult the juniors rating should be altered according to the result but the adult's rating should remain unchanged.

What is to stop well meaning adults throwing their games to juniors leading to inflation of junior ratings and skewing of tournament results? In fact, just the disincentive is sufficient to cause skewing of ratings and tournament cross tables and juniors in the middle to lower of the cross tables can be playing for a junior prize while the adult's only incentive to preserve rating points.

Miguel
02-03-2008, 03:42 PM
What does "euchring" mean?

Rincewind
02-03-2008, 04:01 PM
Euchring (http://www.thefreedictionary.com/euchring) It is a bit like "diddling".

Kevin Bonham
02-03-2008, 07:57 PM
I have a radical suggestion. That where a junior plays an adult the juniors rating should be altered according to the result but the adult's rating should remain unchanged.

A very similar suggestion has already been made by PhilD707. His suggestion was that where an improving novice player of any age played an experienced player, only the allegedly improving alleged novice's rating would be affected.

I commented on this suggestion as follows:

Phil's suggested fix would be inflationary as established players' ratings would only be affected by games against players of stable or declining strength except in the relatively rare cases where a well established player made an improvement. Games against players of declining strength would result in point gains for the established player which they would seldom lose against improving players since their games against improving players would generally not be rated.


Even worse the improving juniors will tend to cause average ratings to fall over a period of time.

There is neither any empirical nor any theoretical reason to believe this is the case in general. In some circumstances it can happen locally but these can be fixed on a region-by-region basis where the evidence is strong enough.

The Glicko-2 system's non-zero-sum nature helps. In most cases where an upcoming junior defeats and experienced adult, the winner gains far more points than the loser loses.

Miguel
02-03-2008, 09:22 PM
Euchring (http://www.thefreedictionary.com/euchring) It is a bit like "diddling".
Cheers.

pax
04-03-2008, 12:23 PM
As any adult can tell you juniors tend to be seriously underrated.
Some juniors are underrated. Most are pretty accurately rated, especially the top ones. Many, however are improving - this not the same as them being underrated (underrated implies that the rating system is not correctly reflecting their actual strength), but can have a similar outcome - performances which exceed their previous ratings.

It is far too drastic a measure to stop rating adults who play juniors. In some places, this would halve or worse the number of games which are rated for the adult players. This reduction could easily be more damaging to the correctness of the adult ratings than any effect of improving juniors.

Ian Rout
04-03-2008, 02:37 PM
In fact the proposal is not especially radical at all, it is just a simplistic version of how the system already works. The differences are

- the system weights an inexperienced player, rather than a junior, less heavily (which is much the same thing if you count a junior as U/18, but for the purpose of the exercise what is a junior? - do you go overnight from Doesn't Count to Counts at 20? 18? 15 years 4 months 17 days?)

- the weighting phases in with more games played rather than jumps from zero to one at the stroke of midnight.

Vlad
04-03-2008, 02:44 PM
Some juniors are underrated. Most are pretty accurately rated, especially the top ones. Many, however are improving - this not the same as them being underrated (underrated implies that the rating system is not correctly reflecting their actual strength), but can have a similar outcome - performances which exceed their previous ratings.

It is far too drastic a measure to stop rating adults who play juniors. In some places, this would halve or worse the number of games which are rated for the adult players. This reduction could easily be more damaging to the correctness of the adult ratings than any effect of improving juniors.

I think the division is different. There are kids who regularly play in the adult tournaments and there are kids who do not. It is true that the first group generally stronger because of selection biase and experience. However, when the two groups meet in the Australian junior competitions, the first group donate a lot of rating points to the second group. The effect is extreme in the under 12 section. Many of the top seeds (with ratings 1000+) lost more than 100 rating points each. Imagine if they were adults, there would be a lot of noise.:)

Vlad
04-03-2008, 03:25 PM
I believe there is a merit in the David' suggestion, which probably requires a small modification. For rating purposes the weighting on the games between adullts and juniors should be high for juniors and low for adults. I understand that this suggestion is probably already implemented, but I think that the differential between two groups should be made even bigger.

Now if 1500 adult looses to 1100 junior, he will loose about 50 rating points. I believe it is just toooooo muuuuuch.:)

Denis_Jessop
04-03-2008, 07:03 PM
I think the division is different. There are kids who regularly play in the adult tournaments and there are kids who do not. It is true that the first group generally stronger because of selection biase and experience. However, when the two groups meet in the Australian junior competitions, the first group donate a lot of rating points to the second group. The effect is extreme in the under 12 section. Many of the top seeds (with ratings 1000+) lost more than 100 rating points each. Imagine if they were adults, there would be a lot of noise.:)

This is right. The juniors who regularly play against adults and especially the juniors who get a high rating are just as badly off as adults in the so-called under-rated juniors situation.

Where juniors don't regularly play against adults and so don't get an ACF rating, or a very reliable one, is there anything that can be done about it without distorting the ACF ratings? Not having detailed knowledge of the system, I simply ask as I don't know the answer.

DJ

frog
14-03-2008, 09:53 AM
Hi All,

Kevin mentioned in post 5 that local adjustments to ratings can be made in regions where juniors have made overall decline to ratings.

South Australia has had two such adjustments over the last 15 years where everyones ratings have been adjusted upwards - this happens where you have more or less a fixed pool of rated players with the juniors over some time taking rating points of strong players.

regards