One Screen System

[canuck]

Been doing this for about a month and am way ahead of the game

Screen 4

Your raw contenders are any horse with a winning prob of 9% or greater

Of these raw contenders-add the PER and CLA figures

The top four numbers are your final contenders-in case of ties preference given to superior PER

Heres the kicker--do not bet anything under 5-1

Sunday was a particularly good day for this method--got a little payback at my home track WO--the 11th race made my week

WO11

Raw contenders 1 4 6 8 9

1 91 94 185
4 82 93 175
6 91 94 185
8 84 95 179
9 81 96 177

We immediately toss the 4 who doesnt make the top 4 list
A cursory glance at this list would suggest the 1 and 6 stand head and shoulders above the other two. For some reason the public ignored the 6 and he went postward at odds of 14-1.
He managed to prevail over the 1 and paid a handsome $30.20

These prices do not happen all the time, but just in the last week have seen it happen enough to believe there is merit in its profitability

[Glen]

Good stuff Canuck. The Prb% stuff is real strong. Especially with a few added factors. PER and CLA are very probably the two strongest non Pace Factors go go along with HTR Consensus, PSCAN, TRN, and Jky. Do you happen do have the Stardust program? It is available on the old Tomcat.exe. It adds the PER and CLA number together so you don't have to do it manually. If you don't let me know and I'll email it to you from my home machine.

[George]

Yeah, not too shabby Canuck. Just used your play to get a nice $21 winner at LAD Race 2.
Also used him with the early horse #3 in a 9,3,all 9,all,3 tri key to get a few bucks short of $1000.

Thanks for the hint :)

[Glen]

Nice hit George, just finished watching the race myself. Surprised they let the 9 go off at such nice odds.

[Glen]

Speaking of LAD...Race 3 #11 (38-1 now). Gimmie him a chance to finish ITM....errr, last.

[Glen]

Rat Droppings, broke last from the field and finished finished 1 length from the winner at 45-1. Back to the salt mines...

[Donnie]

Rat Droppings...? Was that the name of the horse, or what you shouted as he came up 1 length short? :D

[Glen]

Boom, another Ev1 at LAD wins from off the pace for $18.80. Bring out the moola and play LAD Ev.

[Glen]

Sorry to slice and dice your thread again Canuck but another Ev1 Race 8 @ LAD paid $18.20.

[canuck]

George--glad you hit a nice one
If you have access to multiple tracks your looosing days should be fewer than your winning ones
Works real good at Bel and SA and Kee--so so at WO and FE
A real eye opener for me was how many top rated horses were ignored
And also how many 3rd and 4th rated nags got the money
Some days I would feast on 5 6 7-1 shots--other days one $30 mule shows up to make my day
Also not afraid to bet against odds on horses at all

Glen--slice away my man-and go get em at LAD-(is that south of the Mason Dixon line??)

[Cliff]

Glen,

Would you mind emailing me the old Stardust program? Had been looking at Canuck's method. Had not figured out how to program the items yet.

Thanks,
Cliff

cliffhartnitt@home.com

BTW, nice hits ya'll.

[MtKen]

Canuck:

Good idea on your play---There is a horse in the 8th @Mnr--post time 9:55 EST--10/1 ml--that fits thisprofile. Interesting thing is that it is a 2 & 3yo race. The 8 is the 2yo & is getting 10 lbs on the field going 1 mile. It's also a mdn in a nw1 race.

I was looking at this race cuz I got another live horse possibly in the 5 @ 10/1 also. Hit a hot spot play & is a 1st time router who is #2 ped, & the j/t are 40% on the dirt. Might be a good tri race with these 2...

Also,while the fav has the rail the 2nd fav is hung out in the 10 post (6%) & another contender is in the 9 post---I luv to see the contenders go 3 or 4 wide around that 1st turn.
I'm pretty sure Mnr starts right by the 1st turn in these 8f's.

Glad you picked TUP for tomorrow---been getting their simulcast up here for years. It pays to look for the contentious races & let the bankroll fly.

Good hits guys!!! Think I might look at the same thing with TLC3 & the MikeD #....

[Glen]

I'm all for needle sharing. If anybody else wants the old stardust program let me know. BTW, would like to say I bet those LAD horses but I didn't...just watched...

[George]

Glen.........like you didn't bet the ev's. After a good hit usually try to hold on to it by not betting until next day. That way the dollars become "mine". At least until Tuesday!.

The longshot/ev phenomenen usually occurs every year in mid-meet at LAD. Early in meet most trainers still have their better stock at LS. Cole Norman and Jim Gaston have so many horses they send their lesser (closers?) stock ahead. Norman's horses are pounded at the window early in meet and win more often than not. Mid-meet the better horses (early speed?) start coming in from a lot of tracks and surprises happen. In a few more weeks many trainers will start moving the better horses to Florida and New Orleans and back to Cole Norman time then :)

Don't know if their is a real "early-bias" or not. Doubt that much time or money is spent on track maintenance so a bias may be in place. LAD is one of those tracks that is being killed by nearby casino gambling. Last two times I have been there was not more than 50 people in the place. Don't see how they pay the light bill, let alone a maintenance crew. Think simulcasting is all thats keeping them alive. Sad, used to be one of the premeir tracks of the south.

[Bernie]

Glen,

I would also appreciate a copy of the Stardust program. Thanks.

aniruddhan@yahoo.com

[Cliff]

George,

I've also heard that they change their purse structure right after the Super Derby. Seems they lower them considerably. I haven't confirmed this and am not sure what the logic is, but, alot of good horses end up shipping to Arlington Park to catch the end of their meet and then back down to the Fair Grounds. Know for a fact that Sam David, Jr. does this.

BTW, I was at LaD in the late summer and the place was deserted. Very depressing after spending many a weekend there fighting the crowds driving over from Texas. They should get slots in 2002 and maybe that will liven up the place.

Cliff

[canuck]

Played hooky from w**k this afternoon (fri)
My current fav track-Kee-was kind enough to card a race (9th) that provided me with a nifty Screen 4 play

Horses with prob 9% or greater
1 3 4 6 9 11

1 92 99 191
3 89 96 185
4 94 99 193
6 90 98 188
9 91 99 190
11 94 102 196

Horses to make the cut (top 4)

1 4 9 11

The 9 at 18-1 was a serious overlay even though his numbers in PER were slightly inferior to the other 3 who made the cut. His CLA numbers were solid however. He won and paid $38.80

[tbrown]

Trying a variation on this method - using the Prob% to get the contenders then using
velocity to get the winners.
BC Distaff:
5 contenders 1/2/4//6/12

Using Mode 2 paceline (my favorite) the winner is 511 on fractions.
Nice price. To bad I am trying this on paper and used Pratt to bet with!

[HarryH]

I ran some dB tests on several variations of this approach over a 10000 race database covering late May- Mid Aug '01. My best variation did have a positive ROI (about3%) for all tracks. There were about a dozen tracks with 20%+ roi's, mostly the smaller tracks, although DMR and SAR were >15% (truncated season for DMR & SAR). Also ran a spot check on BEL and SAX since Sept season started: ROI of -30%, not good! When I get the time, I will run a forward statistic on the top 20 tracks. .... Possibilities, but not the "Holy Grail"; midway thru the stats, the running ROI dropped to about -38%, but then finished >3%.
Harry H

[canuck]

Harry-a couple of questions

1) Did you restrict testing to horses whose odds were greater than 5-1?
2)Could you run one that restricts horses whose odds are greater than 7-1 and 10-1?

Thanks

[Glen]

+3 ROI is very promising. With more "whittleing" that fig can increase to at least +10 ROI.

[Bruce]

You might have to be careful here, Glen. Didn't Dave Schwartz say the more "whittling" you do, the greater chance the method won't stand up in the long run?

bruce

[HarryH]

Hi canuck,
I used MLO as the odds arbiter (5/1 min) ; I have noticed when testing other systems that the win% drops faster than the profits rise, ==> lower ROI. I will re-run at 7/1 and 10/1
As I remember it, I also tried using post-time odds as filter, but that was not effective (also hard to impliment in practice)

Some of the variations I looked at were betting single MLO>5:1 in top 4 (per+cla) values
and the same only if it was top rF1 (ev1 value) These were very restrictive (fewer plays) and had slightly neg ROI

Harry H

[canuck]

Harry--have found that in real life the MLO is reasonably accurate--but countless times horses with 7/2-9/2 MLO will pay $15-$20 even more!!!
The only drawback to playing this method is consistency--it is very streaky and you have to have serious discipline and resolve to play only at acceptable odds
I have also found that it is rare if you restrict play to horses >5-1 the odds fluctuate detrimentally.
One thing I am sure about is that it really sucks at AP and WO-and performs well at BEL and SA and KEE--so to improve the bottom line I think you may have to be track specific.

[MtKen]

Canuck:

Have you read KM's newsletters about where the longshots reside? I also usually avoid George's K1 favorites or use my horses in exotics with them. I wish I could figure out some way to remove races with strong favorites from my db & test them seperately. Pick your spots & you've probably got a strong play.

[HarryH]

Just finished the additional cases; it took me awhile just to get my analysis program back to its optimum configuration as before. Good News!! Using top 4 contenders as outlined above, picking the single Hi Odd contender with rF1 = 1 (top Ev1 horse) and "Hi Odd" defined as MLO> x:1 ===>
for x=3.5, win%=15% and ROI=10.3% 13 tracks with ROI >20%!
x=5.0 win%=13% and ROI=-2% 8 tracks >20% ROI
This is with no restriction on post time odds. It is definitely a spot play (about 1 in 10 races playable) This was based on a 10,000+ race dataset from 5/25/01 -8/22/01 using "almost" all available tracks. It will be interesting to see if these stats hold up over the year and with different tracks. The good tracks for x=3.5 included a few major tracks (DMR,SAR, APX), and
I believe there were about 20-25 tracks with ROI>0
HarryH

[HarryH]

More Good News: using nJTWpc (Jockey-Trainer Win %) as a filter, for nJTWpc>20%,
the ROI jumps to 29%, with sprints slightly better than routes. Win % at about 18%
Bet rate drops to about 3%, so it is really a spot play. Just remember, as you reduce the number of playable races, the statistics become less reliable because there are fewer samples

[canuck]

Harry-thanks for all the input--ev1 is my fav angle and will be especially vigilant when it creeps into my top 4
Would you mind divulging which tracks were responsible for 20%> roi--surprised to hear that AP was good--guess ev1 as a filter would have saved me a lotta grief!!!
Also good to hear that it surfaces 10% of the time(hi contender plus ev1)
Ditto for any odds-it can get boring looking at a toteboard all day--would love to just bet em and forget em LOL

[canuck]

Harry-another question-ev1 on PL5????

[HarryH]

Yes, Ev1 on PL5. I can give you the "Cream of the Crop" tracks for Min MLO=7/2 and JTWpc>15% ; overall ROI was 28% (38% for sprints)
(also results must have at least 15 plays {arbitrary choice} to make this list):

Track EMD FEX FLX FPX HST MTH BEL DMR
ROI% 39% 58% 42% 98% 53% 90% 93% 143%
Win% 12% 22 16 23 18 16 20 24

I also ran spot tests for PHA and SAX for all races in Oct. There were less than 10 plays for each, with only a single winner; not enough data to really get an accurate picture

I would continue to follow this play forward for a while before commiting "Big Bucks". Or start with a $2 bet to keep your interest and let the "accumulated profits" dictate how large a bet should be

P.S. for this setup, the plays showed 640 out of 14675 races (about 4%)

HarryH

[canuck]

Harry-noticed you live in LA--did you happen to notice the downhill turf(6.5f) 2nd race at SA today? There was a scratch (8) and if I had brought my laptop to the racebook I would have cashed a juicy mutuel.Why? The eventual winner did not qualify to make the raw contenders list because his prob was <9%.However due to the scratch his original 7% turned to 9%

SA2 Raw contenders AFTER scratches
1 2 3 6 7 9

1 91 94 185
2 96 96 192
3 91 93 184
6 92 93 185
7 91 96 187
9 98 97 195

The 3 is immediately tossed and the 6 makes the cut due to superior PER
The 6 won and paid $45.20

[Glen]

My 2 cents regarding using this for Ev...

Ev and CLA rule for longshots
Ev and good PER suck for longshots
Ev and poor last finish rule=Bad PER
Ev and low Prb Rule for longshots

To simplify the Ev play....
Look for Ev1
Good CLA
Poor last finish
Poor Prb

Then you should have a good spot play.

[Glen]

Using Records in the Modeler....Check out the CLA of the Ev1 winners from the last 20 dirt races at Houston Downs.

Win records HOU 020 races Paceline mode 1
Date R# D/S Par Horse Qp RS Ev Lv VEL CLA (K) HTR PScan_r $Pay MLr
------------------------------------------------------------------------------
03Nov 10 6.0D 090 Win 3 S x x x 7 6 4 -4.5 7 43 7
03Nov 09 7.0D 096 Win 1 S 8 1 2 3 4 4 -0.2 2 39 4
03Nov 08 8.5T 096 Win 4 S 4 8 6 4 3 6 -2.0 4 10 3
03Nov 07 6.0D 098 Win 5 P 1 3 2 1 1 1 0.0 1 04 2
03Nov 06 8.0D 090 Win 2 S 2 7 3 7 4 4 -3.4 3 12 2
03Nov 05 8.0D 094 Win 7 E 5 2 3 6 3 3 -5.9 6 07 5
03Nov 04 8.0D 087 Win 7 *E 1 5 3 1 2 2 -2.1 3 04 2
03Nov 03 6.0D 099 Win 7 *E 1 2 2 1 1 1 0.0 1 03 1
03Nov 02 8.0D 096 Win 4 P 5 3 5 4 5 5 -6.6 4 12 5
03Nov 01 8.0D 089 Win 5 P 2 4 3 1 2 2 0.0 1 09 4
02Nov 10 8.0T 092 Win 5 P 3 9 8 1 5 5 -1.1 4 08 6
02Nov 09 8.0D 091 Win 5 S 1 4 1 1 1 1 -0.6 2 05 2
02Nov 08 5.5D 098 Win 7 *E 1 2 2 3 1 1 0.0 1 07 1
02Nov 07 6.0D 097 Win 3 P 7 1 2 2 2 3 0.0 1 30 4
02Nov 06 5.5D 096 Win 3 P 4 4 4 6 4 4 -2.7 6 15 2
02Nov 05 8.0D 096 Win 1 S 9 8 9 4 9 9 -2.9 5 29 8
02Nov 04 8.0D 093 Win 6 *E 1 3 1 2 1 1 -1.8 3 03 1
02Nov 03 8.0D 093 Win 6 *F 1 8 3 3 6 5 -1.1 4 31 5
02Nov 02 5.5D 084 Win 3 *E 1 4 4 2 1 2 0.0 1 05 1


Now check out the your favorite track and see if it holds up.

[HarryH]

One interesting (to me anyway) thing: By using nCLASS alone (no nPER) and picking
rF!=1 in top 4 nCLASS values with MLO>3:1 and only MDN, MDNClaim & STR sprints (whew!!)
I get a win rate of 45% and ROI of 168%!!! sounds good, until you realize that over 14000 total races there were only 11 plays and 4 winners in 3mos. of play!! Talk about your "spot" plays
Harry H

[canuck]

The method picked the highest priced winner I have yet to see today (Nov 8) in the 5th at Aqu

Raw contenders 3 4 7 11--thats it!!!

3 95 96 192
4 95 98 193
7 92 96 188
11 97 97 194

The 4 and 11 went off at big odds-the 11 prevailed and paid $59.50
I was impressed that the value odds has him as a solid bet at odds of 7-1 or greater--talk about a serious overlay....!

[MtKen]

Hope you had it Canuck--George posted it too as a spot play of his--wish I would've seen it!

[Bruce]

Canuck,
Are you using PL 5 to get the results here?

Bruce

[George]

Bruce........in case Canuck is busy thought I would answer. The numbers he is using are static numbers in HTR. They will be the same regardless of paceline selected. PER, CLA, and PRB are calulated internally. Ken Massa in the past has described in various places the general algorithm for each, but I don't remember the exact details.