Couple of Variant Questions

[BitPlayer]

TGJB -

I have a couple of questions sparked by the BC threads.  I apologize in advance if you\'ve already answered them and I missed the answers.

1.  At one point in the threads you said: \"Adjustments of a point or so in either direction from where you think the day is (even assuming you think it did not change) are nothing-- the underlying data is not nearly accurate enough to support the position that what you have come up with is exact race to race (wind readings, wind formula, slight changes in track speed . . . .\"  That being the case, why don\'t you confine your adjustments to whole numbers, rather than using fractions?  (Note:  I don\'t read the famed \"other board.\"  I know there have been posts on this board from time to time about use of fractions in figures, but those posts often refer to posts on the \"other board,\" leaving me in the dark.)

2.  SP suggested that events like watering the track ought to have a relatively consistent effect on track speed for a given set of weather conditions.  I don\'t think you responded to that suggestion.  Do you collect data to confirm that your variant changes following such events are relatively consistent?

Respectfully,

BitPlayer

[TGJB]

Bit-- we use fractions for the same reason we don\'t round the figures off to the nearest 10 points-- they are more usable this way. Otherwise a 15 1/2 would appear the same as a 16 1/4-- both as a 16.

Whatever the degree of accuracy-- let\'s say a point for argument\'s sake-- it\'s that either way from whatever figure you see. So we\'re giving you the best starting place possible.

I\'m going to make time to give Steve a long answer today or tomorrow, because there\'s a lot of meat on those bones. I think everyone will have a better idea of what we are talking about afterward. But I think you (and anyone who has not already done so) should start by viewing the \"Changing Track Speeds\" presentation in the archive section. It covers a lot of what we will be talking about, and shows the scientific studies that back it up.

[BitPlayer]

TGJB -

I think you misunderstood my rounding point.  I wasn\'t suggesting that you round all the figures to the nearest whole number, only that the adjustment should perhaps be a whole number.  If the figure was a fraction before the adjustment, it would remain a fraction afterwards.  I\'m not trying to be critical, just trying to understand.  Given what you posted earlier about the amount of imprecision inherent in the data, how do you decide that the adjustment should be plus or minus a quarter point (see the BC Juvenile Fillies and Classic), rather than just leaving it at zero?  The argument against using fractional adjustments is that you\'re artificially making the data smoother than it really is.

Respectfully,

BitPlayer

[TGJB]

I\'m not sure I get the question. We add or subtract both because we use the past data to come up with current figures, and because we recognize that the underlying data is inherently somewhat innacurate. I don\'t make small adjustments to individual horses, just to the race as a whole-- just as I make major adjustments to the race as a whole, and for the same reasons. It would not make sense to add 3 points to get it close to where it needs to be, but not add another half point to get it tight.

If you want to see what I mean, play around with the BC races we posted, adding or subtracting a half point to ALL the horses in a race.

And what would be the rationale to thinking that adding or subtracting whole points (an artificial unit of measurement) would be good in some way?

[BitPlayer]

TGJB -

I understand that all horses in a race rise and fall together.  I think I misinterpreted your earlier post that I quoted in my original question.  If you think that adding or subtracting a quarter or half point to the variant makes the figures from the race more accurate, then I would agree that you are right to make those adjustments.

Respectfully,

BitPlayer

[TGJB]

Bit-- you don\'t have to be so respectful. It makes me feel old, like when a kid calls me \"mister\". I tell them mister is my father.

The basic idea is that even though we accept the inherent limitations of accuracy, we are trying to make things as accurate as possible. If you start letting half points slide, it will eventually loosen up your data base, and you will have trouble putting things together. The same thing goes for not using all the elements (wind, ground, and weight, for example)-- other skills aside, the best you can do is to get a race generally accurate.

And the same goes for using an average as a variant when conditions are not the same-- you end up with a sort-of generally correct variant, within a few points. That\'s the degree of accuracy those who use pars get-- and it was plenty good enough back in the days when 99% of the bettors weren\'t using speed figures with any basis in  reality. In other words, until the advent of the Racing Times.

Damn you, George White.

\"Damn you, George White.\"

That was funny.

IMO, that\'s what makes some of the more subjective elements of trip handicapping like projecting the pace, position, and ground loss, observing bias, and other qualitative things valuable these days. They really can\'t be packaged and sold because the adjustments you make in your betting are often part of the real time handicapping and betting process.