Paceline -- F1, Cramer-RS-POS

[andicap]

Jim Cramer from HDW selects pacelines for determining a first fraction projection by taking the horses' best f1 in his last 10 races. (RAW, not adjusted).

What are people's feelings about this?.

Obviously Cramer's done a ton of research into this so it's not to be taken lightly. And I understand the rationale. Even an off-form horse can go very fast for 2f (sprint) or 4f (route), and in a heavy pressured race an early horse could be expected to go his maximum speed out of the gate (assuming no bad luck, etc.)

But most horses have one PL that is totally out of whack -- that you know there's no way the horse will re-create that line. It is too good. Or the horse was hurt since that line, or turned 7, or got a bad trainer, etc.

[cj]

If you are telling me he doesn't take into account the track, I think it is not a very good system. Pimlico and Churchill, the same horse can run the 1/4 2 seconds slower at Pimlico, and run exactly the same speed!

[sjk]

If you're looking for another opinion, it's hard to imagine that any serious research couldn't do better than this.

[GameTheory]

Here's what Cramer says in the RS-POS info sheet:

The "Pos" part of Rs_Pos™ is calculated by predicting how fast each horse can run. If all of
the horses in the race broke out of the gate and ran as fast as they possibly could, what
would their positions be after a 1/4 mile? Without getting into a discussion about all of the
methods we have tested to determine this, we continue to use the best 1/4 mile a horse has
actually run in the last 10 races. The correlation between this ranking and finish position is
nearly perfect.

If you use the best 1/4 time (B1/4) for each horse in the last 10 races, the horse that ranks 1
will win more races than the horse that ranks 2, who will win more than the horse that ranks
3, who will win more than the horse that ranks 4, etc. The horse that ranks 7th will most likely
be 7th at the first call and will most likely finish 7th. This fact gives us a powerful tool to use in
our handicapping.

Remember - this is raw speed - it is not adjusted in any manner. Horses that can run fast
have demonstrated that they can run fast! The big error here is to try to make some type of
adjustment. Of course we know that when horses move from a very fast track to a very slow
track, they slow down, the point is they all don't slow down the same amount. Conversely, to
assume that horses speed up when moving from a slow track to a fast track is incorrect - not
all of them do. We do know that "E" horses are more affected than "S" horses, but is very difficult
to say, accurately, by how much.

[sjk]

The true test is how often you are suprised by who makes the race lead and by who makes an easy lead. A length or two makes the difference between a horse that gets clear and has an easy trip and one who contests the lead and tires.

A careful pace analysis will be right a good percentage of the time as to who makes the lead and how contested it is likely to be.

Perhaps the method proposed can do that. I would be a sceptic, but since I have not tried it I can only wish good luck to those who use it.

[sjk]

andicap,

In case you are still interested in this question from yesterday I figured I would give the computer some exercise to see if it would validate the assertion in GameTheory's quote from Cramer.

According to my data it is quite true that the top ranked horse on this measure wins the highest percent of races, the second ranked horse wins the second highest percent of races, etc.

You might then ask whether this helps at all in terms of ROI. To my surprise the same was true for return: The top ranked horses as a group showed the highest return (-9.9%) and each successive ranking group showed a lower ROI than the one before.

I would still assert that a more careful analysis of pace will do much better, but there is more to the simple best f1 in last 10 races measure than I would have guessed.

[andicap]

thanks,
not surprised since jim cramer has the biggest database of almost anyone.
I too had a hard time believing it was true and agree with you that there has to be a better way.

[formula_2002]

I would appreciate one or more of you placing the value on what you are talking about.
By value I meam $$$ profit.

A thorough breakdown would be fine.




Thanks
Joe M

[Tom]

Crammer is not using this paceline to handicap the horse - just to identify the rank of best F1 races. He has a rigid formula for assigning running style as well.
Then you line the horses up by running style and rank of F1, such as

P1
E2
E3
P4
P5
E6
S7
S8

In this match up, the P1 horse is worth looking at closely. The E6 is probably headed to pace problems. What this is saying is that the E6 horse has a best F1 that is slower than 5 other horses.

The thing about this tool, and it is only a tool, not the full handicapping process, is that is is fully automatic - no judgement required, and can be tested in a db.
Crammer has found this to be a useful tool and so do the many people who pay for the service each month.

The idea of raw times is unaceptable to many players, but let me say that at Finger Lakes, we have seen bottom claimers run fractions of :22 :45 and allowance horse run :21.3 :44.4 at times. Does this mean that top stakes horses would be running :20 and :42 splits?
Case in point, the speedball Groovey ran here and won at the heiht of this career and recorded fractions no faster than he did at Belmont or any other track.

[Jeep]

IMHO if you're going to rate a horse on it's time at the first quarter, you should use the best time where the horse finished decently not just the best time overall. What if it finished 10th by 12 lks. I'd rather use it's 2nd or 3rd best time where it finished less than 6 lks behind at the finish.

[kitts]

Jim Cramer is meticulous in his computing of RS/POS data. I worked with Jim and he spent hours each day with this. He bases his research and calculations by racing circuits and the finish position of the horses. He recalculated each circuit every 90 days. The resulting data is solid, very solid.

[so.cal.fan]

I have used the RS_POS method.......here in So. Cal. It works on dirt pretty good, not on turf.
I adjusted it as follows......
I use his method of determining the Running Style. As for Position:
I use it with distance races on dirt only........I use one of the last 3 or 4 pace lines.......for an equal or better class only.
I have found this fairly accurate, however it has to be used with other handicapping and/or physicallity on race day.
I don't like to have too long a memory........pace lines back too far are meaningless in my opinion.


Note: let me clarify......I only use this method on distance races.....as sprints are way too closely contested in regards to early speed, here in So. Cal.
I will use a paceline from a sprint, of better class, if it is recent.
I also do not use the 1/4 mile time......just the half mile fraction.

[Tom]

There is a method from Fred Davis years ago to identify early speed that uswed only positions - first call position at today's distance or shorted. Add the two lowest for each horse. So a horse running 3 and 2 as its best early position would rate 5.
The nice thing here is that it is sometimes not obvious and the prices can be good.