About Real Tennis - Rules of Real Tennis - Tennis and the Chase - Handicaps

How do Real Tennis handicaps work?
This question surfaces every couple of years when players are baffled by the apparent black art behind handicap adjustments.

Handicapping is only ever an approximation of relative skill level. Match results will vary, such that if A has a better handicap than B, it doesn’t mean that player A will always beat player B – but over a number of matches, A would be expected to win more often than not. Playing off handicap evens things up and gives both players scope to prove they are better than their rating.

A handicap difference of 6 points is assumed to mean that a game played from a 15–0 start will be equally likely to be won by either player. If those players started instead from 0-0, then the better player would be expected to win 2 out of 3 games. This is the statistical basis of the likelihoods of different outcomes based on actual handicaps and those played.

The handicap difference between two players is reviewed in light of a match result between them, to decide whether that difference is probably accurate, or needs adjustment. If a change is needed, then each player’s handicap is altered by the same amount, one favourably, the other not. Thus your handicap is based on many results, and the more results, the more accurate it is likely to be.

To determine the size of any handicap adjustment, the system considers the match result in the context of the outcome that was expected, which depends on any difference between the players’ actual handicap difference and the handicap difference at which the match was played.

The system works out the games won by each player, adding an extra game for each set won e.g. a score of 6-4, 3-6, 3-5 would translate to 13-16, so the first player won 13/29 or 44.8%. We'll refer to these as 'outcome points' to distinguish from handicap points and points in each game!

The expected result depends on whether the handicap played matches the real handicap difference. If it does match then the ‘effective’ handicap difference is 0 and the expected result is 50% each. If not, then the expected result varies according to the difference between actual and played handicaps. If the real handicap difference is 2 points but the match is played off 0-0, then the weaker player is under-compensated by 2 points, and so is expected to win only approximately 43.3% of the games. The approximate proportion of games the weaker player is expected to win for each effective handicap difference value is shown in the table:

The statistical challenge is how much deviation from the expected result is acceptable before it merits adjustment of the handicaps. Current practice recognises three outcomes:
• a draw i.e. the result hits the expected figure plus or minus around 8% (so approximately 42-58% for Level)
• a small win/loss, i.e. outside the draw range, which moves the handicaps by +&- 0.2
• a big win/loss, which moves the handicaps by +&- 0.6, and is applied to extreme cases of the above which occur less often than 1 in 10.
This mechanism does mean that the 'better' player needs to win well enough to justify the handicap difference - it can turn out that though the better player wins the match, it records as a handicap draw or even a loss because the winning margin was not sufficient.
NB The percentages given above are for guidance only and are not absolute. The system in fact uses a table of threshold values measured in games and sets won, according to the effective handicap difference played (as described above) and the total number of outcome points (as above). For each combination of these, a set of threshold values define the ranges of outcome points which constitute a big loss, loss, draw, win or big win. The table is attached below, and as an example, if there are 14 points for, say, a correctly handicapped 8-5 win, the entry for 0 difference and 14 points shows that 0, 1, 2 or 3 games means a big loss, 4 or 5 a loss, 6, 7 or 8 a draw, 9 or 10 a win, and 11, 12, 13 or 14 a big win. So the 8-5 result, reflected as 9 points to 5, is a win/loss.
Click here for a table of the handicap thresholds
Other features which might affect you and should be worth looking out for include:
• home court advantage, to reflect it being less likely that a visitor will do well (so a visitor's handicap is deemed to be higher than shown)
• 'volatile' players whose handicaps are changing so quickly that special steps are needed to adjust them faster by using stable players as yardsticks (doesn’t necessarilly mean they are temperamental!)
• weightings, to increase the size of handicap adjustments after more important club competition and National League matches
• pro adjustment, so Alan and Andrew can manually disarm bandits improving too fast for the system!