View Single Post
Old 01-21-2012, 07:02 PM   #82
BJennet
Grade 1
 
Join Date: Jul 2008
Posts: 311
Actual loss is even worse

Quote:
Originally Posted by mikesal57 View Post
lets do better....


here the numbers so far for the winners at odds 4-1 or better for NP contenders

E1 = 70 at odds 4 or better. Bet $140.00, won 7 for $90.80....loss of $49.20

E2 = 109 at odds 4 or better. Bet $218.00, won 9 for $148.70....loss of $69.30

L1 = 81 at odds 4 or better. Bet $162.00, won 7 for $102.80...loss of $59.20

L2 = 113 at odds 4 or better. Bet $226.00, won 10 for $253.60...profits of $27.60

T = 30 at odds 4 or better. Bet $60.00. won 2 for $56.60....loss of $3.60

made 403 bets at $2.00 for a total of $806.00, 35 bets won for a total of $ 652.50 for a loss of $ 153.50
Hi Mike,

Sorry I haven't posted much here, but I've been somewhat tied up lately. I don't know if you saw my post on the analysis of NP higher-odds winners from Bill Lyster's 507-race survey, but it might have made the ROI-loss here less of a surprise. And your actual loss here would have been worse than 19%, because, if you're using only odds as a filter, you have to use all of the NP contenders above your chose level (here, 4-1). So you have to assume that you weren't only using the winner, which is what you do here, but that you were using at least one and possibly two more horses in every race. Much bigger loss. This is due to NP's lack of rank-ordering, as I've said before, a huge weakness of this method.

As I said in my post on Bill Lyster's suvey, NP's hit rate deteriorates sharply as the odds rise. The average win probability for all four NP contenders combined, even at the level of 7-1 and above was 8% (2% for each contender) rather than it's average of 74%. That's pretty close to what you got here. If you are interested in using an odds-filter, I suggested starting with 12-1 or 15-1, at the lowest.

As Jeff Platt has said, the more filters any factor or method requires to produce a profit (and most produce no profit to begin with) the less robust it is. Based on what we've seen, I think we should assume that NP is a relatively weak method (in terms of ROI), that can only produce a profit in very limited cirumstances. The MNR angle is a good example of that.

I am leaning toward a belief that exactas are more the way to go with NP, due to the compounding effect of the odds, and the fact that the crowd seems to have more difficulty assessing correct exacta probabilities compared to win.

Again, thanks for your continuing work.

Cheers,

B Jennet
BJennet is offline   Reply With Quote