Steady State vs Punk-Eek Posted on April 7, 2019 at 02:43:26 PM by Craig G
I'll go with punk-eek on this one and predict that an 8 consecutive point run-out occurs more often than the Skiena-type model (representing steady state) would expect.
We might imagine that each player - for a variety of reasons - has a range of intensity / skill levels that are not day-to-day constant. In reality, more likely a continuum, but we can simplify things by quantizing it into A-B-C-D-E levels, where most of the time the player will be in the C state.
I think that in general, an extremely dominant player who wins about 61% of his points will have an overall win rate of about 1 in 4. For him we might get something like: (crude guesstimates)
A - 80+%
B - 68%
C - 61%
D - 56%
E - 45%
With this kind of reasoning, we predict that most of the amazing run-outs will occur when he is in the A-state. Or maybe B.
This cuts both ways, so that if A-state correlates with more run-outs, then in observing a run-out (and not thru lucky points) we can also predict the player is more likely to be in the A-state.
We could make a wild-ass prediction to that effect wrt Diego. Ie, we predict (just based on that one game - crazy?) that he is probably at the top of his game lately.
1 - last night Diego did win again in his next doubles game
2 - arbitrary base period of last 24 perfs shows an even and sustained 1 in 4.6 win rate
3 - Urbieta not that great over 24 perfs, but en fuego beyond question in last 40 games with 8 wins
Incidentally, I noticed that Erik ran-out 4 out of 5 singles games just about a week ago. More great support for punk-eek philosophy. (And what about that Ibon surge?)
But do you think having Erik play to 9 points would have made much difference? I don't. So he could have popped out several 8-consecutives, right there.
The bottom line to all of this is that if you use some means to estimate, or even tabulate, a player's point-winning percentage, then use that to predict run-out rate, it will produce a significantly lower figure than the quantum model, which expects bursts of higher performance levels.
So I predict that an examination of all games in jai-alai history would show more run-outs than you might normally expect.