Analyzing my SAR results
 

I've been a horseplayer about as long as I've been on this board. One consistent piece of advice is to keep records. I've played Saratoga pretty aggressively this meet, and broke down my 300 wagers over 18 race days thus far, several of which are small part wheels counted as one wager.

Total credits are 3% lower than total debits. I topped out at 154% ROI before falling to -29%, mostly on the weak cards last Thursday and Friday—bet size went up and hit rate went down. Curious what you think of the results breakdown.

By race type, I was profitable in higher quality races: allowance (34%), optional claim (22%), and stakes (19%). Stakes account for a third of my handle. I fared poorly with maidens, especially claimers. This is kind of intuitive, although I felt more confident about MCL than my results show. Also aligns with my recent slide on those poor cards.

The most surprising result is that I have a 23% return with fillies & mares, but -24% on males. I didn't expect a substantial difference by sex. Is there a logical explanation, or just a consequence of small sample size?

By age, I was profitable with all 3 year olds at 39% and F&M 3+ at 59%. Maybe 3 year olds hit a sweet spot of limited but not scant information to project forward for improvement.

Another curious result is 63% return on the inner turf, but -37% on the firm main turf track (although I got a small profit on the brief run of "good" main turf bringing the total to -31%). I nearly broke even on a fast main track. In the mud, I went broke with dirt runners but only lost half my money on off-the-turf runners.

Distance is tough to assess. I'm 21% on routes, but mostly from the inner turf, which doesn't host sprints. I also scored 34% on the 9f dirt routes. I tanked at 8.5f, on both turf courses, -58%. Probably just an anomaly?

For dirt sprints, I did O.K. short at 5.5f and 6f, but poorly longer at 6.5f and 7f. I bet a lot of 5.5f main turf races, losing at a 20% clip. Oddly, I did well with claiming sprints on the main turf, as well as MSW dirt sprints and the cursed 8.5f on main turf, categories in which I otherwise fared poorly.

I mainly wager win, place, and exacta, with the rare trifecta. A classic would be 2x win (occasionally dutching two runners), 4x place, and a few 1x exactas. My exactas almost always include runners with complementary styles. I'll play savers with or without favorites, swings against the favorite, and chalk enhancers when I have opinions underneath. I limit the spread to six combos, usually less, and equalize the probable payouts. But those exactas are killing me at -36%. My place wagers prevail with 27% ROI, and win eked out 2%. I must be catching prices to place that aren't running with their exacta companions on top. The win dutches may be deflating the ROI vs. place.

My place and exacta handle is about the same, about a third more on win. I'm emphasizing win/place because I'm getting an ADW promotion that gives rebates equivalent to 10% takeout, so my net is a bit higher. That's good, because my inclination is to play exactas and trifectas, which apparently I don't do well at. Not sure I'd do better if I focused more on exotics rather than using them to enhance win/place. The latter seems the "correct" approach.

Just wanted to see if anyone had any insight about my results, especially how I might interpret my success with fillies and mares, the inner turf, place bets, and my failures with the main turf, long sprints (vs. short) on the dirt, and exactas. Or is it all just too small a sample to be meaningful? Also let me know if there's anything in particular I should cross tabulate.

Sorry for the long post, hard to make it short.

I guess my first question would be, is this Saratoga meet representative of past results as well? I know the phrase "sample size" dominates most conversations on this board, but I don't think much can be gleaned over data from one meet...just my two cents.

You might try adding a data point designed to capture why you made each bet in the first place.

I use a series of letter codes for this.

But a number (1, 2, 3, 4, 5, etc.) would work as well.

I've found I need (at a minimum) several hundred tickets sitting in a database.

But once I have that -- I've found that (for me) analyzing the data in terms of why bets were made does a better job of pointing out strengths and weaknesses in my game than does other data points such as surface, dist, class level, etc.

I've given considerable thought as to why this is so.

In my own case the codes I use are data points that describe something subjective -- something unique to me as a player.

This falls into the category of thick data.

Which dovetails with the discussion that took place a while back in this thread:
http://www.paceadvantage.com/forum/s...d.php?t=139762




-jp

.

Good discussion!!

I cant add as I have taken off considerable time and dont have enough bets to make any analysis at Saratoga.

For me though its clear I am doing much better with dirt sprints.

Jeff makes a great point. We all have certain preferences under specific circumstances. The data that you shared is absolutely necessary, but it means very little unless you can identify the conditions or circumstances that led to your decisions. I've played this game for nearly 40 years now, and I wish that I could tell you that it is possible to eliminate the qualitative components of this game. I was a consistent loser for years, until I started consciously taking note of my degree of emotional intelligence, or lack thereof.

Been thinking about my comments regarding the use of qualitative data. I want to be perfectly clear that it's not possible for ME to only use quantitative info when wagering. I know for a fact that some on this board use only numerical data to beat the game...it's just not for me.

When I work with players, we always begin by asking these questions (in this order):

1. How did your contenders do vs. your non-contenders?
In other words, are the horses you call non-contenders losing big money?

If not, then your contender selection needs work.


2. Are the non-contenders you select that ultimately go off at 3/1 and below losing big money? (target should be $net of $1.20 or below.)

If not, then you do not know the difference between a good favorite and a bad one. That means your assessment of low odds contenders needs work.


3. Are the contenders you bet returning significantly more money than the ones you don't bet?

If not, then your value assessment of the horses themselves needs work.


4. Are the races where you bet a lot returning more money than the races where you bet a little?

If not, then your race selection/race valuation method needs work.


IMHO, this is how you analyze your results.


Regards,
Dave Schwartz

This advice makes sense, thank you. I've already alluded to different reasons I wager, so I think I can implement that fairly painlessly. Mr. Schwartz's questions take a bit more work to answer, but shouldn't be too arduous to chronicle the contender choices and results.

Seems a bit cheap in hindsight trying to make hay out of the data I had, never mind that it took considerable effort to tally. I was anticipating answers more like "well, the unique thing about the inner turf at Saratoga is X, and your handicapping must be picking up on that."

Good stuff here, though. Thanks again.

It would be useful to look at the ROI on various types of bets.

When I first started out I made money on place, show, and exacta boxes.

I did not do well on win bets, straight exactas, keyed exactas, or trifectas.

I started boxing all my exacta wagers and my ROI went up. I was a good at picking contenders and the feedback helped me improve my betting style.

Here is one of two methods I’m using to analyze my data.
My play requires a factor >=0.30. That factor is determine using a fixed formula
The play also requires but one such play per race.
Currently I have reviewed 451 races.
182 of those races contains a factor of 0.30.
77 of those races contain but one horse whose factor is >=0.30.
See the attachment for the $1 returns for the win and place bet for each of the indicated segments.
Note, most plays are often the favorite. The odds ranges are rather small.
If the play continues for be profitable after 500 races, I'll present a statistical analysis commonly use by Dr Burton P. Fabricand PH.D.

THIS CHART INDICATES THE HIGHER THE FACTOR THE GREATER THE WIN BET RETURN, WHILE THE LOWER THE FACTOR, THE GREATER THE PLACE BET RETURN

THIS CHART INDICATES MUCH BETTER $1 RETURNS WHEN THE FACTOR SIZE FALLS WITHIN THE RANGE >=.30 AND <=.45

What does ".30" mean, please?

0.30 what?

It represent a factor which is derived from my formula based on the live 1 minuet to 2 minute tote board data.

Have I missed an explanation of this factor?