FR3

[DOUG]

Would someone try this one for me?

Tried it on the robot on about 2000 races various tracks and it tested pretty good.
not too many plays though. all dist, sur age etc.


pl-5

k2,3,4,5.

fr3= 1
layoff < = 90
work > = 80


Thanks,

Doug

[Victor]

Total Bets 6667
WIN PLACE SHOW
Total Amount Bet 13334.00 13334.00 13334.00
Wins 1324 2550 3584
Pct. 19.86% 38.25% 53.76%
Amount Won 11801.60 11415.50 11559.80
Profit/Loss -1532.40 -1918.50 -1774.20
Pct Profit Loss -11.49% -14.39% -13.31%
Avg. Payout 8.91 4.48 3.23

[Victor]

That covers 75,117 races.

[DOUG]

Thanks Victor.

Not nearly as good as my little test.

Doug

[Victor]

That's the way it is Doug.

I have not found too many Big samples that hold up, not to say there aren't many! out there.

In fact, I have one or two.

[Victor]

I think "the Cane" among others have stated that F3 is difficult to hang one's hat on.

I agree except if you're also picking your own lines, then F3 can be very good.

[Glen]

Think AP1/Vel1 and Fr3 rank 1 hold up pretty well.

[AlanLN]

Doug,
Here is results from what I have for the month of November, all tracks, abit over 4000 races. Looks close to breaking even.



ITEM Plays Wins WN% W-ROI WP% Long-%W-High P-ROI
--------------------------------------------------------
MyPlay 00399 0092 23% 0.97 41% 0010 11 $32 0.85

K-1 00000 0000 0000
K-2 00123 0040 33% 1.17 50% 0003 08 $32
K-3 00109 0024 22% 0.86 45% 0001 04 $17
K-4 00107 0016 15% 0.62 29% 0002 13 $16
K-5 00060 0012 20% 1.38 37% 0004 33 $27
K-6 00000 0000 0000
K-7 00000 0000 0000
K-8 00000 0000 0000
K-9 00000 0000 0000

K110 00003 0002 67% 1.10 67% 0000 00

$ 00005 0001 20% 1.58 60% 0001 10 $16
$$ 00123 0020 16% 1.16 32% 0008 40 $27

FR1-1 00008 0002 25% 0.88 38% 0000 00 $10
FR1-2 00020 0008 40% 1.25 60% 0000 00 $9

E/P-1 00005 0001 20% 0.45 40% 0000 00 $5
E/P-2 00018 0004 22% 0.61 44% 0000 00 $7

A/P-1 00116 0033 28% 0.87 47% 0000 00 $11
A/P-2 00086 0016 19% 0.69 45% 0000 00 $17

S/P-1 00289 0072 25% 0.92 45% 0002 03 $27
S/P-2 00071 0015 21% 1.29 38% 0002 13 $32

L/P-1 00239 0064 27% 1.09 46% 0003 05 $27
L/P-2 00096 0019 20% 0.95 36% 0002 11 $32

FR3-1 00399 0092 23% 0.97 41% 0005 05 $32
FR3-2 00000 0000

VEL-1 00167 0044 26% 0.91 47% 0001 02 $22
VEL-2 00121 0022 18% 0.77 40% 0001 05 $27

C90-1 00123 0036 29% 0.95 50% 0001 03 $22
C90-2 00105 0023 22% 0.78 43% 0000 00 $17

HTR-1 00115 0036 31% 0.91 50% 0000 00 $12
HTR-2 00113 0023 20% 0.79 43% 0001 04 $27

WRK-1 00192 0040 21% 0.79 36% 0002 05 $27
WRK-2 00104 0028 27% 1.16 51% 0000 00 $17
WK80+ 00399 0092 23% 0.97 41% 0005 05 $32
WK85+ 00121 0030 25% 0.99 42% 0001 03 $26

TRN-1 00057 0017 30% 1.06 51% 0000 00 $13
TRN-2 00080 0021 26% 1.20 44% 0001 05 $27
T400+ 00012 0003 25% 0.82 50% 0000 00 $12

JKY-1 00065 0020 31% 0.79 52% 0000 00 $8
JKY-2 00052 0014 27% 1.10 38% 0001 07 $32
J330+ 00058 0018 31% 0.87 53% 0000 00 $8

PED-1 00051 0015 29% 1.16 55% 0001 07 $22
PED-2 00060 0013 22% 0.66 37% 0000 00 $12
P450+ 00098 0022 22% 0.84 45% 0001 05 $22

MLO-1 00084 0030 36% 0.89 60% 0000 00 $8
MLO-2 00102 0023 23% 0.78 38% 0000 00 $9

PScan <= 3.0 0081 27% 1.02 46% 0005 06 $27
----------------------------------------------------------------------
Races Tested: 4069 $15+ Wins= 24% Tourn-Rating= 67 PL-5

Track= ALL Alw/Stk Clm Msw Mcl 2yr 3yr 3up/4up Females Male
Purse<$10k Purse>=$10k Vi 15-26 Vi 27-32 Vi 33-37 Vi 38-50 All-Surfaces
4.5f-5.0f 5.5f 6.0f 6.5f 7.0f-7.5f 8.0f 8.5f 9.0f-9.5f 10.0f-14.0f
K2 K3 K4 K5
Fr3= 1 Layoff< 90 Wk 80+

[hurrikane]

I think if you start out with one of two things it will increase your w% and ROI. Either use a contender filter(Cont = 1 is ok), and use consistent horses..no 2yo/3yo only races.

Then break it out by class. Fr3 seems unusally better with better class horses and as was brought to my attention lately
by the time or year. Seems ev1/2 diminishes in the winter. Either sealed tracks, smaller fields, or KM California software doesn't like cold weather.

Anouther theory I haven't thought of until now. Many of my spots were dragging the last couple of months but have gone nuts since the election.
Maybe I need to sort tracks by red tracks and blue tracks. :D

Mikee...can you write a parser for that! :eek:

[MikeDee]

hk I'll write the parser now all you need to do is to get Ron to put state color in download

[Victor]

rFR3=1
nQFives>4

Total Bets 2488
WIN PLACE SHOW
Total Amount Bet 4976.00 4976.00 4976.00
Wins 397 751 1074
Pct. 15.96% 30.18% 43.17%
Amount Won 3574.00 3975.60 4051.30
Profit/Loss -1402.00 -1000.40 -924.70
Pct Profit Loss -28.18% -20.10% -18.58%
Avg. Payout 9.00 5.29 3.77

[Victor]

I for one am shocked!

I thought with all those speed horses in the race the FR3 horse would benefit.

What about nQFive=0?

rFR3=1

Total Bets 2530
WIN PLACE SHOW
Total Amount Bet 5060.00 5060.00 5060.00
Wins 494 908 1266
Pct. 19.53% 35.89% 50.04%
Amount Won 4628.50 4227.60 4313.00
Profit/Loss -431.50 -832.40 -747.00
Pct Profit Loss -8.53% -16.45% -14.76%
Avg. Payout 9.37 4.66 3.41

Shocking.

[Victor]

SELECT ALL_HX4.tTRK, ALL_HX4.tDATE, ALL_HX4.nRACE, ALL_HX4.tPGM, ALL_HX4.tHOR, ALL_HX4.xFIN, ALL_HX4.xWIN, ALL_HX4.xPLA, ALL_HX4.xSHO, ALL_HX4.rFR1, ALL_HX4.rKrat, ALL_HX4.rFR3, ALL_HX4.nWKSC, ALL_HX4.nLAY, ALL_HX4.nQFives
FROM ALL_HX4
WHERE (((ALL_HX4.rKrat)=2 Or (ALL_HX4.rKrat)=3 Or (ALL_HX4.rKrat)=4 Or (ALL_HX4.rKrat)=5) AND ((ALL_HX4.rFR3)=1) AND ((ALL_HX4.nWKSC)>=80) AND ((ALL_HX4.nLAY)<=90) AND ((ALL_HX4.nQFives)=0));

Total Bets 190
WIN PLACE SHOW
Total Amount Bet 380.00 380.00 380.00
Wins 45 81 117
Pct. 23.68% 42.63% 61.58%
Amount Won 395.80 366.50 369.50
Profit/Loss 15.80 -13.50 -10.50
Pct Profit Loss 4.16% -3.55% -2.76%
Avg. Payout 8.80 4.52 3.16

[hurrikane]

interesting V-
but it eliminated to many races for me to make a value judgment.

one thing i'll toss to you though..your queries will run faster with fewer Or statements. try using either = between 2 and 5 or >1 and <6. it's less thinking for the box.

[Victor]

Thanks hurrikane,

my database only includes the nQFives beginning July 10th, 2004

Using Doug's query but with nQFives=0 and rKrat=1, we get something for the XFAV to think about:

Total Bets 72
WIN PLACE SHOW
Total Amount Bet 144.00 144.00 144.00
Wins 32 49 55
Pct. 44.44% 68.06% 76.39%
Amount Won 148.80 155.20 145.10
Profit/Loss 4.80 11.20 1.10
Pct Profit Loss 3.33% 7.78% 0.76%
Avg. Payout 4.65 3.17 2.64

[Cliff]

What date version has the nQfives export?

Thanks,
Cliff

[Donnie]

Cliff---
Since Ken first introduced this at the seminar in July I would assume any version with a July or later date on the Start Up screen.

[tbrown]

F3 is just the last fraction and might include too many horses that are too far back. Maybe using LP=1, which brings turn time into the equation, might be better with a lot of Q5's? With all that front end speed, maybe you need to ensure your horse is making a move in the second fraction as well.

[Huguenot]

Mike Pizzolla adds in the number of beaten lengths to the final fraction for this late speed number as a way of discarding horses who make a big late move from far back and finish 5 lengths back. Is there a way of trying that?

[Donnie]

Huguenot--
The numbers are in the HX5 export to do the calculations from the beaten lengths to either the stretch call beaten lengthes or the second call beaten lengthes. Then the LP numbers are in HX4 I believe. Let us know what you find!! :D

[km]

I have tested the Pizzola method as well as the James Quinn Turf method with FR3. Neither works as well as S/P.

S/P is the rating of choice for a balance between decent early position and a strong final fraction (Fr3). But the win rate, as is for Pizz and Quinn methods, is in the low 20's% and ROI in the mid 80's range.

Fr3 by itself destroys all of these methods for longshot identification. But the win rate is very low.

S/P, Pizz and Quinn methods seem like good common sense approaches to finding a horse than can remain in contact with its field and then rush home with that strong Fr3. But isn't that what we and the public are looking for with every racehorse we bet? It is common sense that defeats us in the ROI chase.

Strong FR3 horses are frustrating because they usually meander sluggishly toward the rear of the pack and just don't move up most of the time. The key is to find a "live" one that will make the move - instead of trying to formulate a better rating.

To find "live" Fr3 horses experiment with the Wk rating, positive Trn and Jky switches, or those with higher consistency ratings (good "LIVE" or "ITM" percentages).