"Number of good speed ratings in the last 6 races?"

[Bill Cullen]

Is anyone familiar with the concept of "the number of good speed ratings in the last 6 races?"

I came across this concept in "Longshots" (abridged edition) by Mike Nunamaker. You take the horse with the highest speed rating last out. Then you subtract 5. Then for each horse you count the number of times the horse met or exceeded this value in its last six races. According to the statistical table in the book by Nunamaker, the impact values peaked at horses having three "good speed ratings" in their last six races.

In other words there was no advantage, impact-value wise, by having 4, 5, or even 6 good speed ratings in a horse's last six races.

I should note the above stats by Mike Nunamker were for horses 10/1 or higher.

I was curious if anyone knew if the above finding also holds true for horses going off at less than 10/1.

Very intriguing finding!

Thanks,

Bill C

[Bill Cullen]

The silence on this post speaks volumes...

Bill C

[dav4463]

I like this kind of analysis. I don't have much to add, but will definitely take a look at it.

I do a similar thing with Beyers. I look at the two best of last three and add them to the best figure from races 4-10 back as long as there isn't a one-year layoff in there. I use the total as a guide to pointing out some longshots that may hit the board at a price.

[Bill Cullen]

Dav,

Thanks for your response.

Bill C

[Overlay]

If I'm understanding the metric correctly, you're saying that impact values drop off for longshots with four or more speed ratings that are within five points of the highest in the field in their last six races. It would seem that the more such races that there are in a horse's record, the more consistent and competitive the horse would be.

Therefore, if such horses are still at high odds despite this, there must be other negative handicapping aspects that are resulting both in their being bypassed in the betting, and in their failure to win races. The lower the horses' odds go, the less likely it would be that there would be such holes in their records.

So I would expect that, for a lower-odds horse, a greater number of good speed ratings in its last six races would make it more competitive, rather than less, and would not be characterized by a drop-off in impact values.

[Bill Cullen]

If I'm understanding the metric correctly, you're saying that impact values drop off for longshots with four or more speed ratings that are within five points of the highest in the field in their last six races. It would seem that the more such races that there are in a horse's record, the more consistent and competitive the horse would be.

Therefore, if such horses are still at high odds despite this, there must be other negative handicapping aspects that are resulting both in their being bypassed in the betting, and in their failure to win races. The lower the horses' odds go, the less likely it would be that there would be such holes in their records.

So I would expect that, for a lower-odds horse, a greater number of good speed ratings in its last six races would make it more competitive, rather than less, and would not be characterized by a drop-off in impact values.

Thanks for your response, Overlay.

You understood the metric correctly (at least as I read and understood it).

I tend to agree with your extrapolations to lower priced horses, but my real puzzlement is this: I can not understand why the more negative attributes you theorize on the 4+ horses at high odds would suddenly show up in this single metric. I know of NO known correlations with any of hundreds other known stats or tendencies about thoroughbreds with this data. The sample sizes for the metric are in the tens of thousands.

Even more puzzling, this changes from race to race, being driven by whatever horse happens to be the last-out speed rating leader.

This metric is, for me, until other explanations prove otherwise, the most potentially powerful, intriguing and revoltionary insight into the nuances of thoroughbred handicapping that I have come across in at least the last 15 years.

Best,

Bill C

[andicap]

Too many better than average races the horse is due for a "correction." I hate to use the term "bounce," but as Horatio Luro used to say, paraphrasing here, you can only squeeze the lemon so much.

[ryesteve]

I can offer another explanation as to why 4+ might work out no better than 3: if the horse has that many "good" figs, it's possible that the target fig was very attainable for this group of horses; in other words, a lot of horses in the race might have 4+, which would dilute the impact value.

An alternative to this metric might be to use a fixed par fig dependent on the class of the race, rather than a "moving target" based on the recent races of the horses that were entered.

[jasperson]

If I'm understanding the metric correctly, you're saying that impact values drop off for longshots with four or more speed ratings that are within five points of the highest in the field in their last six races. It would seem that the more such races that there are in a horse's record, the more consistent and competitive the horse would be.

Therefore, if such horses are still at high odds despite this, there must be other negative handicapping aspects that are resulting both in their being bypassed in the betting, and in their failure to win races. The lower the horses' odds go, the less likely it would be that there would be such holes in their records.

So I would expect that, for a lower-odds horse, a greater number of good speed ratings in its last six races would make it more competitive, rather than less, and would not be characterized by a drop-off in impact values.
Possibly the cause of have 5 or 6 speed ratings over the rating last race -5 is because his speed rating last race is too low
case speed last 49
1=49 2=50 3=60 4=70 5=50 6=69 7=55 8=51 9=55 10 65
In this case all speeds are above 44. but I wouldn't like to bet on this horse
Change his last race to 66 speed rating and he has only three and the horse look like a better bet.
Jack

[rufus999]

Sounds like too much work for this Bozo... I use the Beyer figs or any other published ratings for grouping purposes primarily. When I dope out a race I lend more importance to race flow and conditioning. Is the horse live and if so, does the race set up well for it.

rufus

[ryesteve]

Possibly the cause of have 5 or 6 speed ratings over the rating last race -5 is because his speed rating last race is too lowThe "target" is based on the best last fig in the field, not the last fig of the individual horse.

[jasperson]

The "target" is based on the best last fig in the field, not the last fig of the individual horse.
That makes it more clear

[rokitman]

Is anyone familiar with the concept of "the number of good speed ratings in the last 6 races?"

I came across this concept in "Longshots" (abridged edition) by Mike Nunamaker. You take the horse with the highest speed rating last out. Then you subtract 5. Then for each horse you count the number of times the horse met or exceeded this value in its last six races. According to the statistical table in the book by Nunamaker, the impact values peaked at horses having three "good speed ratings" in their last six races.

In other words there was no advantage, impact-value wise, by having 4, 5, or even 6 good speed ratings in a horse's last six races.

I should note the above stats by Mike Nunamker were for horses 10/1 or higher.

I was curious if anyone knew if the above finding also holds true for horses going off at less than 10/1.

Very intriguing finding!

Thanks,

Bill CThe only way this horse is over 10/1 is if all the good numbers are coming from "back form" which is impossible if it has 5 or 6/6 and near impossible and not far enough back to matter if it is 4/ 6 , not to mention the horse has shown a helluva lot of competitive recent races (assuming no major lay-off) so there is not much reason to get over 10/1. As you well know, nothing is given away in this game. And, obviously, 1 and 2 of 6 didn't do well either. Sounds like the very convenient wishful-thinking kind of data mining that blows up in one's face. No way this held up in any meaningful way. The maximum pool percentage here was around 7.5%. A very unstable vein of your date mine.

[Bill Cullen]

The only way this horse is over 10/1 is if all the good numbers are coming from "back form" which is impossible if it has 5 or 6/6 and near impossible and not far enough back to matter if it is 4/ 6 , not to mention the horse has shown a helluva lot of competitive recent races (assuming no major lay-off) so there is not much reason to get over 10/1. As you well know, nothing is given away in this game. And, obviously, 1 and 2 of 6 didn't do well either. Sounds like the very convenient wishful-thinking kind of data mining that blows up in one's face. No way this held up in any meaningful way. The maximum pool percentage here was around 7.5%. A very unstable vein of your date mine.

I didn't mine any data: it's from a table in Mike Nunamaker's book, "Longshots"
(the abridged edition). What I have been trying to do is understand why the data is distributed the way it is. It looks totally out of whack from anything I've previously known about thoroughbreds. My implicit assumption has been that Mike Nunamaker has correctly crunched his numbers. Maybe he didn't, or maybe the tabular data was incorretcly printed. I have sent queries to Mr. Nunamker to find out

The only wishful thinking I can be accused of is hoping that the numbers are for real. No doubt there's a very valid and very unexciting reason that will explain away the numbers.

To date, I have yet to see an adequate explanation that explains the data one way or the other.

Thanks for your response.

Bill C