Handicapping

Long enshrined on Wednesday nights at Cambridge, handicaps had a dark history. Happily this is no longer the case. An explanation of how we used to arrive at and apply handicaps is pasted below this thoroughly modern method now in use.

Dynamic Handicapping - the New Method

XG ratings are produced from scores recorded when you play X-Clubs deals (as we do). An updated XG rating is assessed weekly taking into consideration your XG, your partner's XG, and the XG's of your opponents in every session you played. Your XG rating is a component in calculating an Expected Score for you with any partner in any strength of field every time you play. If you exceed Expectations then your XG should increase slightly next week. And vice versa should you fall short.

When you play an X-Clubs session Compass scoring now applies this same methodology to your raw score to get your handicapped result for the session. Your "handicap" is simply the difference between your Raw Score and your Expected Score - how well did you play against yourself? This is not the same as the averaging of your combined XG's as it allows for the strength of your opposition. A comprehensive explanation is given on the Compass website.

Dynamic Handicapping can (and should) be set for all events in the Club Calendar. It is List 19 - by NG/XG Grade. Set and Forget - no more messing around with ...

 

The Old Method (for anyone who wants to see what hoops we used to have to jump through)

Our scoring system, Compass, records both raw scores and handicaps for all sessions. Everyone thus has a basic history.

Calculation

Compass is able to look back into last season to gather sufficient data if we tell it we wish to consider the last fifteen sessions. We choose to do fifteen in order to more accurately assess current form without the baggage of what happened eighteen months ago dragging things down. Compass can produce all the detail about who played with whom on which date and what score they got to back up its calculation which is:

The highest and the lowest scores of however many results it found are dropped and the average of the rest gets subtracted from 50 to get a fix on how much above or below "average" each player has been scoring. (The scores of any session average out at 50%.) This figure is further averaged with the player's old handicap to tie it in with at least some previous history - much in the same way that NZB ratings are calculated.. This is known as the "Tauranga Method"

Some manual editing now takes place. Compass will not update player's handicaps if it finds less than five previous results. We go with what it says for nine or more results (the "greens") but need to research the "oranges" (six to eight) and the "reds" (zero to five). To do this we:

Use the same parameters to produce a set of "handicaps" for Mondays and Thursdays even though those sessions are unhandicapped. These results are used only for assessing the marginals (reds and oranges) in this handicapping exercise because we have established a relationship of Wednesday handicaps = Monday handicaps minus 2 or Thursday handicaps minus 1. Using this formula we are able to assign handicaps to the "reds" and cross-check the handicaps calculated for the "oranges".

All this is followed up with an eye-balling by the club captain before updating the database. It's quite a lot of work and we don't like to do it too often - four times a year, in fact. Some other clubs assign a handicap for the whole season, whilst others just go with whatever Compass comes up with. We opt to take more care as handicapping is such a minefield.

Application

We have traditionally used the system setting that applies the average of the two players' handicaps to the raw score when a pair plays together. Bob Fearn, the writer of Compass, recently provided the ability to apply a user-chosen percentage of the sum of the two handicaps. In this light, the method we currently use might be regarded as adding the handicaps of a pair and then taking 50% of the sum. We can now experiment with changing that to, say, 70% or 35% of the sum.

An example: Player A, who learnt last year, has a handicap of +5.2 and plays with a more experienced player B who is on -3.4. In one session they get a raw score of 52.4%. Player C, another newbie, also has a handicap of +5.2 and plays with a classmate player D who is on +4.4. Their raw score of 44.6%. Under the current scheme the average of A and B's handicaps is applied to their raw score. Here 5.2 + (- 3.4) = 1.8 which averages to 0.9 ... this gets added to their raw score of 52.4 to produce a handicap score of 53.3%. The other pair's handicap score is 44.6 + half of (5.2 + 4.4) which comes to 44.6 + 4.8 = 49.4%.

If we choose now to take 70% of the combined handicaps the results for the first pair will change to 52.4 + 70% of 1.8 = 53.7%. The second pair will get an adjusted score of 44.6 + 70% of 9.6 which comes to 51.32%. The handicaps, which have been produced fairly scientifically, are having a bigger effect on the raw scores with the higher factor used.

In summary: we are very happy with the way the handicaps are produced and just need to tinker around with how they are applied.