nVBET

[fred4now]

Anybody use nVBET? Not sure what to make of it. I ran this over all tracks and distance.

r(k)=1, nODDS>[nVBET]

Bets 6363
# Wins 911
% Wins6 14.32%
Bet $12,726.00
Win $10,505.00
Profit ($2,221.00)
Win ROI -17.45%
# Place 1927
% Place7 30.28%
Place $9,816.00
Place ROI -22.87%
# Show 2868
% Show8 45.07%
Show $9,417.20
Show ROI -26.00%

[George]

Fredfornow.....not sure what the question is? nVbet is the reciprocal of probability percentage and is a good tool to compare with post time odds. Substitute rTRN=1 for K1 in your query and should get a nice surprise. Filter it a little further by rACL and type race and it gets better.

Used vbet with mlo for a long time as part of a spot play. nK2 greater than 1.0 also tells you that you have an overlay as it is the ratio of vbet and mlo.

Only problem is near post odds are so volatile any more it is difficult to apply the play.

[fred4now]

Thanks for the reply George. I guess my question was that betting these when odds are greater than nVBET is supposedly an overlay but definetly don't show a profit. Why not? Not trying to be picky just curious since they are in HTR and even in TLC.

[George]

Fred4now......thought that was your question. Age old problem with "value" betting. Once you get into "overlay" odds the win rate is so low the payoff isn't there. Real value lies between 4-1 and 20-1 but only on certain factor combinations. Trick is to find the combination.

There are not many and the people that have found them are understandably reluctant to reveal them.

Not touting HSH but that software does a nice job in it's reports of pinpointing the "overlay" odds range for a race-type/factor. You can see at least historically where the hit rate is high enough to maintain a profit through an odds range.

You can do the same thing with HTR db just a lot more work.

[Glen]

Why is it in there? Because it is a tool to help us beat the take...The true take isn't 16% or whatever when betting all horses...More like 25%....

Here is something that shows a profit using Vbet in my dB...

nVBET>=3.5 and <=9
nOFF >=([nvbet]*0.95) And <=([nvbet]*1.25)
Fr1
nD/S=1

252/1590
15.85%
1.08 ROI
$13.71 Avg Mutal

I'm sure with a little bit of salt and pepper, these numbers could look even better...

[George]

Interesting how great minds think alike :)
I use an equation very similar to Glen's with wider vbet range (4.0 - 20) but same "takeout" modifiers (.95 & 1.25). Unfiltered TR1 gets 14.7% and 1.08 roi, very close to Glen return.

Adding a couple more filters brings it up to 18.9% and 1.27 roi.

Am sure there are better paying combo's in there and agree vbet is a useful tool for finding them.

[fred4now]

Again thanks for the reply guys, that answers my question.

[Glen]

Rick,

Kinda weird, but on the limited testing I've done, if the horse's odds are way over the VBET number then it is a true pig. Public very good and it seems to be more than difficult to get a good true overlay...Sad but true...Since I don't get a chance to watch the tote much I haven't looked into it much, but plan on doing so in the near future. Getting ready to "train" for the Orleans tourney and that is one of the things I gotta get a better handle on.

George, thats funny you use the .95 to 1.25 multiplier...It is something that I queried and keep in the back of my brain but haven't had a chance to practice it in real life...

[fred4now]

Glen, guess that was my real point. The line is very fine on what is an overlay and what is a bad bet. Agree the public is really good. I'm going back to my spot plays, if they show positive ROI, I know they are an overlay.

[ERNIE]

Suprised that Glen didn't remember to refer to some of his own work.......(on the old board )
1/13/2001 Glen posted "play the percentages" where he reported on the results of the %Prob. and how often that percentage won......using some 80k races.......his report confirms that it is a solid piece of information...........this report is one of the post that I printed and stored with the newsletters....Thanks Glen....... I think that using the vbet as a contender valididator is important...I.E. like to see my top contenders have a highier %prob than the m/l,trouble is that all morning lines aren't constructed by handicappers.......some tracks use Ms.Cleo....and dat's what me tinks ........... :D :) ;)

[Glen]

Interesting point Ernie...Actually ran a few more test w/ VBET and there are a handful of factors that show a small positve ROI (1.05 to 1.15) using VBET w/ a couple other positive factors...Imagine if one made a list of all those factors and played daily with large win bets, they might be able to make a small living at this using this angle...

[fred4now]

I remembered Glen's post too Ernie, just couldn't find it. I did this one last month over about 57,000 races.

[Glen]

Something amazing that you can add to that is to do nPRB w/ Ev=1 and then nPRB w/ Ev>1...The Win% and ROI will be very revealing...

[Carl]

Nice chart Fred for Now.

Thanks for posting it, enjoyed reading it. Quick question: do ROI numbers stay constant throughout as well?

later,
Carl

[fred4now]

Good question Carl, didn't test for ROI just Win %. Will take a look at Ev1 Glen.

[Carl]

Yeah, I was wondering if there was "a sweet spot" for probability. I doubt if it would be at either extreme of the curve, but maybe there would be some near profitability in the mid-range bunch.

And 57,000 races may be too small a sample to find this anyway. Sounds crazy to say that, especially when you consider Quirin did his landmark studies on a 600 race sample at one curcuit! Now if I could only have back all those dollars I spent chasing "two sprints to a route", invested at compound interest........

later,
Carl

[hdcper]

Was hoping one of you HTR users was going to show how many horses fall in each nPRB% category and the overall ROI.

Just wondering,

hdcper

[Glen]

http://www.homebased2.com/cgi-bin/UB2/UltraBoard.pl?Action=ShowPost&Board=htr01&Post=459&Idle=60&Sort=0&Order=Descend&Page=0&Session=

[Cliff]

Glen,

In your post of July 16th, what does nD/S=1 refer to? I'm not familiar with that terminology. Sorry to ask so much later.....had a note to ask that got buried under papers.

Regards,
Cliff

[Glen]

No prob Cliff. It is terminology form Impact Export. Field 07 is the distance/surface where 1=Dirt Sprint.


http://www.homebased2.com/km/dnl/Impexp.txt

[Cliff]

OK,

Thanks, Glen.....That would explain it as I have not been exporting the Impact data as of yet and learned all the nuances (so little time, so much to learn!).

Good luck to all you guys in the big tourney this weekend!

Later,
Cliff