Beyer Speed Figure question (Belmont)

[joeydb]

Steve,

If you get the chance to ask Andy on a future occasion, I was wondering if the comparison of American Pharoah's Belmont performance versus that of Secretariat would change either quantity from a speed figure perspective.

Andy has said on previous occasions that Secretariat's Belmont would be equivalent to a 138 or 139 Beyer Speed Figure.

We also know now that American Pharoah's figure from Saturday was a 105.

Does this 33 or 34 speed figure point difference make sense with the 2.6 second difference in observed final time? Also, both horses led at all the points of call so the observed time throughout the race is their time, with no lengths necessary to apply as a delta.

Joe

[ateamstupid]

Originally Posted by joeydb

Steve,

If you get the chance to ask Andy on a future occasion, I was wondering if the comparison of American Pharoah's Belmont performance versus that of Secretariat would change either quantity from a speed figure perspective.

Andy has said on previous occasions that Secretariat's Belmont would be equivalent to a 138 or 139 Beyer Speed Figure.

We also know now that American Pharoah's figure from Saturday was a 105.

Does this 33 or 34 speed figure point difference make sense with the 2.6 second difference in observed final time? Also, both horses led at all the points of call so the observed time throughout the race is their time, with no lengths necessary to apply as a delta.

Joe

You're looking at this wrong. Beyers are computed relative to the times of other races that day, not races from past years, so you can't compare times from 42 years apart and try to match up the time differential with Beyer points. It's apples to oranges.

[tector]

Originally Posted by ateamstupid

You're looking at this wrong. Beyers are computed relative to the times of other races that day, not races from past years, so you can't compare times from 42 years apart and try to match up the time differential with Beyer points. It's apples to oranges.

But you can compare the figs and then convert that on the beaten lengths chart. See page 104 here: https://goo.gl/4Ks2Qt

[ateamstupid]

Originally Posted by tector

But you can compare the figs and then convert that on the beaten lengths chart. See page 104 here: https://goo.gl/4Ks2Qt

I understand that, but the races were 42 years apart.

[Indian Charlie]

Originally Posted by ateamstupid

I understand that, but the races were 42 years apart.

It does seem kind of silly, doesn't it? The 42 year thing that is.

[tector]

Either you largely buy into the premise of the figs or you don't (including the retro calculation for 1973). If you do, then comparisons over decades are no more problematic than say comparisons in the same year. YMMV.

[Danzig]

Originally Posted by tector

But you can compare the figs and then convert that on the beaten lengths chart. See page 104 here: https://goo.gl/4Ks2Qt

but it's all still an exercise in trying to decode. and horses can't do math.

[joeydb]

Originally Posted by ateamstupid

You're looking at this wrong. Beyers are computed relative to the times of other races that day, not races from past years, so you can't compare times from 42 years apart and try to match up the time differential with Beyer points. It's apples to oranges.

OK, but since Beyers are compared to each other all the time for multiple contestants, which is the point of speed handicapping, there must be some correlation between the actual speed and the speed figure. We don't require that the Beyers to be compared be earned on the same day at the same track.

I was under the impression that one of the advantages of Beyer Speed Figures over the old speed index approach with the track variant was that the difference in track condition was already rolled into that single number so that comparisons would then be easier.

Acknowledging that there is likely to be a huge difference in track conditions over 42 years, is the difference between the speed figure values over the same distance by the respective horses believable, even after taking into account a very large track condition difference?

[Danzig]

Originally Posted by ateamstupid

You're looking at this wrong. Beyers are computed relative to the times of other races that day, not races from past years, so you can't compare times from 42 years apart and try to match up the time differential with Beyer points. It's apples to oranges.

thanks for this.
just like all the pearl clutching and hand wringing over times with all the races, this year and years past.
we know a track can change one hour to the next, but people still want to look back at how horse a, b or c stacked up. records are all well and good...
but i recall when a certain horse won the derby in a slow time...and then the preakness. and then he won the belmont, and slew is now in the hall of fame.

[tector]

There is no question the track was way fast Saturday. I don't have a problem with the fig.

[ateamstupid]

I think we're having two different arguments. Who the hell is disputing the utility of a beaten-lengths conversion? It is obviously essential to figure-making, as is the fact that times and figures from two different days on two different racetracks have zero correlation.

The original question was "Does this 33 or 34 speed figure point difference make sense with the 2.6 second difference in observed final time?" This question completely misses the point of speed figures. The 2.6 second difference and the 33-34 Beyer difference don't need to "make sense" together. They're two completely independent times and figures from two different days.

The time differential from two races 42 years apart has nothing to do with the Beyer differential. The Beyers, after factoring in pars, are only made in relation to the other races that day. American Pharoah could have run the race in 2:30 and still gotten a 105 if the other times on Saturday dictated that. Again, the time differential between AP's Belmont and Secretariat's Belmont are immaterial to the speed figure discussion.