K1 Four Star

[George]

After MLO the most consistent percentage winner in HTR factors is the K factor.
Across all tracks and races the top 2 ranked K horses will win over 50 percent.
Spent a lot of time and some of my money trying to use K factor during my first
3 months using HTR. Problem is horse that gets K=1 rank is very obvious to
the public. Moved on to early speed, trainers, and jockeys and left K for others.

About a year ago became intrigued with trying to identify "weak" or vulnerable
favorites and reanalyzed K looking for factors that might indicate failure.
In the process came up with a play that tells me not only when I have a weak
favorite but also when I have a strong one. Once done a few times the play is
very quick to do on a card using the HTR screens.

One caveat. I never play MDN, MCL, or 2 yr olds using factors so those type races
are excluded from this play. Ignoring those kinds of races have wagered on this
going forward on over 2000 races. Have won 55% with 8% roi. Have not kept score
but am sure have made more $$ by knowing when I can safely throw out the top K.

Rules:
Screen 0
1) K rank = 1
2) Days off < 41 (could extend on ALW races but I don't)
3) JKY rank < 6
4) HTR # rank < 4
Screen 7
5) ACL rank < 4
6) ACL # > PAR (race par listed in upper right scr 7)

If horse meets all of above go back to screen 0 otherwise go on to next race.
Screen 0 Ratio Calc
Set Pace Line to 1
The K horse is HTR # 1,2, or 3. Sum the HTR #'s for other two HTR ranked horses
and divide by the K horse's HTR #. Jot down PL1 ratio.
Set Pace Line to 5
The K horse is HTR # 1,2, or 3. Sum the HTR #'s for other two HTR ranked horses
and divide by the K horse's HTR #. Jot down PL5 ratio.

If ratio for both PL's of the qualified K horse is < 1.47 the horse WILL win or place.
If one or both ratio's are above 1.46 but below 1.70 must keep in exotics but probably
won't win.
If either ratio is greater than 1.70 Throw him out. If you have another strong play in
the race with good odds, loosen the rubber band on your bankroll.

A nice recent sample race to look at is LAD 08/12/2001 Race 1. The #4 qualifys on all
factors above but ratio's are bad. Horse ran out in a 5 horse race.

[hurrikane]

Hey, thanks George..I have it recorded a little different. I know you keep tweeking it, thanks for the update.

I agree it is very good at finding the weak fav..as well as the strong ones. I use it every day....

[George]

Sorry about the tweaking Hurrikane. The above is how I ended up "qualifying" the K
horse but there are many ways one could do it. Am sure there are better ways than I am
currently using. The important thing is once you have a K that "should" win is his HTR number
overpowering in relation to his two nearest competitors. If not, he is vulnerable.

BOL !!

[Ron]

George;

Thanks for reposting the special situation K1 play so succinctly.

I like to think of it as a five star play.

One star for each of the 5 next qualifications the K1 contender passes on your list.
(If you want to give a star for the fact that it's a K1 to begin with, I guess it could be a 6 star rating)

Anyway; It helps me remember that I need to check and verify 5 other readouts on my K1 contender. Also, in keeping with Kens 10% per star winning probability it works out to be an over 50% play for sure.


Ron

[jackad]

Using this method in a very small sample showed that betting only if MLO is only > = 2/1 improved results. Betting only MLOs between 2/1 and 3/1 proved best of all. Will anyone try this in a db?
Jack

[George]

Jack,
Think your odds analysis on this play is accurate both short and long term. In real world wagering
it is very difficult to play odds ranges on these.
They are the type horses that sit at 2-1 and get banged down to 4-5 after the bell. Conversely, have seen them swing from 3-5 to 2-1 after the bell. Can be frustrating.

To run your analysis is difficult because one needs to keep a PL1 db AND a PL5 db on same data.
I do not currently keep both.

Think Hurrikane does and perhaps he will run some
odds range analysis for us. Like you, I would be very interested in knowing "best" range.

I try not to play one below 6-5 but do get caught with way too much wagered on a 3-5 quite often.

Which is why I like the play better as a "weak" favorite identifier. Then some other idiot has him
at 3-5 and I have "next best" at 8-1.

[jackad]

Findings might be instructive even if test is run with db of either PL1 or PL5. Again in my very small sample, I show essentially the same results if selections qualify in both PL modes or only in one.
Jack

[Bernie]

How about testing with just a single paceline as jackad suggests, as well as filters of MLO>1, MLO>1.5, MLO>2, etc.? Dual pacelines and actual odds would be better, but at least we could see a large sample workout that could conservatively approximate the real world. It might also be able to indicate what tracks this approach works best at - and if there are some where it doesn't do as well as at others. I'm curious to know if George has found anywhere where this approach has not been so good.

[George]

Bernie,
Surprisingly, this play works almost everywhere. Does not do well at PHA but have never found anything that really works there.

Will wait on Hurrikane to check in as I am certain he can easily check the odds ranges using both PL's. He runs these as spot plays every day so his db is set up for both pl's.

The play is statistically better if both PL's are considered. The crosscheck back to PL5 insures that we are not dealing with a current race "flash in the pan".

If Hurrikane is unable to run it I will run against current data for each track and PL. Have
done the test using both PL'S long ago but no longer maintain two db's. Did not filter by odds ranges when I did the test.

[hurrikane]

Sorry folks...I don't keep a full database of Pl1 just run it for the day and then delete the pl1 and build my db on PL5.