Council meeting

[Andrew McHarg]

One of the problems I see in selecting purely on an average, especially a statistically biased one based on a selected sub-set of results, is that players have different results profiles: some will will have great wins against highly rated opponents combined with poor results against lower rated opponents while others are inherently more consistent. They beat those they should beat and loose to those they should loose to if you like. To know who is the better player you need to understand the reasons for the differences.

I agree that players with similar grades may have radically different journeys to getting those grades, but I'm not sure I agree that we can assert one to be stronger than the other purely on preference of one of those records over another. I mean, are we saying here that we prefer a player who consistently plays at a particular level over a player who can occassionally beat very much higher rated players but who also loses to much lower rated players in equal measure? I'd say the pros and cons of each roughly equal each other. We could say that consistency is a valuable quality, and it surely is, but so is the ability to put in an inspired performance. In fact, I'd say the latter probably shows more potential than the former, but does that make this player any stronger at the moment (?); there's probably not much in it. Therefore I'd have to conclude that the only fair way of assessing their relative strengths would be to assume they are roughly equal.

I think a good idea would be to look at which player has improved the most in the last 12 months (i.e. grade rise). It's maybe more credible to suggest that a player who has risen 120 points in the last 12 months is playing better Chess than someone who has only risen by 10 points; if their live grades after subsequently equal. Maybe that could be suggestive of "form"? It would be interesting to analyse the statistics on the grading database to see if the performances of high risers outweight the performances of low risers (or droppers) of equal grade in x games immediately following such a rise/drop.

The alternative to selecting players based on their five best results, selected with hind sight, the selectors could nominate at the start of the season a selection of events for players in the squad to play at least five from to be considered for selection and their performances at these events becomes a major selection criteran.

This not only tests their chess ability but their ability to turn up at a specific event and perform on the day - as they will have to for the event they are hoping to be selected for.

At these events selectors would attended and observe the players and provide coaching after the games and if practical before. This would both help the selectors get to know the players but gets the players partly used to the routine involved in major events.

It is only appropriate for players above a certain age or ability but it is something they do in swimming and it seems to work.

This is a really good idea I think. It sort of combines the selection process with training for the big event. Therefore, it may actually assess who can best handle the pressures of performing when it counts?