Saturday, February 23, 2008

Handicapping: Pulling the Trigger on Value

Well we are at the last part of the series. In part one we looked at why we need to get value. In the second part, we talked about identifying value. This one is about finding and betting your value play.

Where do we find these plays? They could be anywhere, really. Sometimes they are right in front of you. Trainer angles are usually worthwhile. So, let’s find a trainer angle and look at how to recognize it, when to bet it, when not to bet it and when to jump off of it.

Everyone knows Casie Coleman, she was trainer of the year in Canada two years ago, and is a high percentage winner. We will use some of her stats to find an angle (as an exercise) and then make a value betting plan.

We’ll first look at her off claim stats from Dec 1st 2004 to December 1st 2005 at Woodbine. She had 24 fresh claims. She won with only two, and had a horrid UTRS, as well as an ROI of -80%. Unless we want to bet and lose, we should stay away from fresh Casie claims. But we must keep watching, because trainer angles change.

Over the next couple of months - from December 1st 2005 to February 1st 2006 - a funny thing happened. She made 7 claims. She won 2, placed 4 times and was third once. UTRS was 0.651: She did not miss the board with fresh claims. We are now onto something. With a 0.651 off claim UTRS, after a dismal off claim record, we must make a decision because this statistical fact might continue, and this is in our wheelhouse: A high percentage bet.

This is our plan moving forward after recognizing the anomaly: She hits 29% off the claim now, so with fair odds she is a little less than 5-2 odds break-even. We decide that any of her horses that are off the claim, who are over 5-2 odds we will have an edge, and we will bet.

On her next 28 claims at Woodbine she went on to win 12 races, for a 43% hit rate. Her ROI on new claims was a whopping 26%, and many of these horses were 2-1 or over. We were in good shape pulling the trigger on these bets. As you reevaluated her hit rate, you could then lower your fair odds. At the end of the period you might have been betting her at 3-2 or over (you need only over a 40% success rate to make money on 3-2 shots and since she was hitting at over 40%, you got value).

After this period, a couple things happened, she cooled off a bit, and she was overbet. The run with this angle was over.

Ed Bain, in his book and on his website bets what he calls “4+30’s”. This alludes to his automatic bets where if he sees an occurrence happen 4 times and it has at a 30% or more hit rate, it is a bet. When betting these plays, we set the bar for our minimum odds at 5-2, because 5 to 2 is profitable at a 30% hit rate. If he notices something happened 50% of the time, he might bet that horse at say 6 or 7 to 5 as a floor. It is a simple way to quantify value with angles.

Some examples from the World of racing? I’ll try a couple:

Wednesday at Gulfstream in the runners, trainer Kirk Ziadie had one off the claim. He was off 60+ days and he was dropping. In this situation this trainer hits at over a 50% hit rate. He would be a bet at even money or better according to our value line. At post time: He was even money! It was a red light for us, but there was opportunity. In 20/20 hindsight, he won like a 1-5 shot, so even money might have been an overlay; but we don’t care about that.

I am currently running similar models. I have one that has a 29% hit rate and an 8% edge. One was in today. He was even money. I could not pull the trigger on that. I would lose two out of three times and only get paid one out of two times! Red light.

Another angle last night fared better. I had a 9 out of 23 trainer stat (40% hit rate, so 3-2 is our floor), and the horse was 5-2 (28% break-even). Green light.

Another angle, this one for thoroughbred turf racing happened today. I have a stat that is a whopping 47% hit rate and the horse was 3-1! Wow, I'm rich. The only bad part of the story is the horse lost.

The results of the past are not a true indication of the future. We all know that. But by approximating the hit rate of your angle, you can at the very least approximate what odds level you will take to pull the trigger. If your pattern becomes too obvious you have to pretty much drop off the bandwagon as you will have very few green light bets. Seldon Ledford near the end of his career was at that point – he was massively overbet most times off the claim. Kirk Ziadie at Gulfstream is similar.

Hidden positives are the key. We need to find something that happens 20% or more of the time, which is not bet heavily. When we have those we have found our wheelhouse as 5-1 or better it is a play. If something happens one in five times and we are getting 10-1, it is time to swing for the fences. You can not get too low a hit rate though, because of the massive variance. If something happens 4 out of 100 times and the winning horses paid $60 I would pretty much ignore that other than taking an action bet, eventhough it is ROI positive. You need high hit rate angles.

Conversely another thing I watch for is what I said above: We can’t accept 1-2 or 3-5 on an angle we see happens a great deal of time. This is difficult because when you see something win that often you begin to think it is invincible. It is clearly not. When the angle gets to that level, it is time to move onto something else. If you are bridge jumping that is another matter of course. I am purely speaking of spot plays.

I hope this illustrates how I handle angles, or trainer patterns. I think it is pretty sound. First you have to recognize one, like with Casie off the claim; then you evaluate, then approximate fair odds, then pull the trigger based on your perceived hit rate. When the angle goes south, we drop it like a hot potato.

I know what some might say about my betting of these angles, especially trainer ones: That is, we never look at the individual horses and individual races with this and handicap. That is true, but you know what? To me it is irrelevant. If Seldon Ledford claimed a horse and improved it 6 lengths and then another one 8 lengths and another one 4 lengths how could we possibly know which is which? How can we possibly make an accurate fair line with that? When dealing with trainers like that you are in a chasm of the great unknown. You have to take all your passion out of the wagering equation and look at hard numbers. That’s what I like to do anyway. Andy Beyer said supertrainers have soiled the great art of handicapping. He is right.

I am working on this every day. I am trying to find the right bet amounts, and frequency. It is very difficult. If it was easy, I guess the game would not be much fun. It sure is fun trying though, because when you do succeed, the sky is the limit.

Any comments or thoughts?

Note: For you thoroughbred watchers Phil at his play of the day blog has a good handicapping primer on Laurel Park. Well worth reading, so give it a shot if you are interested in playing LRL. He handicapped tomorrow’s card as well. It gave me an idea for a track bias post for harness and I think I will work on that for this week.

Note 2: If this post is of interest and you want to learn more about betting angles, I would suggest Ed's book above. It gives a wealth of information on how he gathers, bets and filters stats. It is not tough to understand at all, and for any player it is a good read. You don't have to be a runner player to use his knowledge. It works with any game of skill.


Phil J in NJ said...

Thanks for the plug!

Interesting that you made this post when you did. Over the last two weeks, I have started tracking the top 50 or so trainers at the Meadowlands and another 20 that I have interest in for one reason or another.

I did this because I have always thought that trainer intent was even more important in harness racing than thoroughbreds. One afternoon at Chester this year, a friend and I were down along the track near the 3rd turn and Sean Bier ( believe, I want to say him or J. Roberts ) pulled up with his mount prior to the filly and mare open and asked if we were gamblers ( No buddy, we heard the hot dogs were good ). To which we replied yes and he told us we could bet his mount today, because they were going to try to win. Sure enough, he did just that. I thought the "try to win" was interesting way of stating it. Not, she's ready today, or we made a change of the driving bit and she should race better, he said we are going to try to win.

I also noticed that often times horses trained by N. Varty were especially live on the board when making their second starts for him. Or that Ross Croghan's charges were often well bet off a qualifier.

Sure enough, it has been quite interesting. With limited samples so far, I've uncovered some real useful things.

R. Croghan wins at 38% off a qualifer and 31% 2nd start off a 14+ day lay off.

Richard Johnson is 6 for 6, 100% with horses in horse barn for the first time.

John McDermott is awesome with new aquisitions.

Nat Varty is as I suspected awesome with horses making their 2nd starts in his barn. ( See Art's Image from a few weeks ago, 20-1 or so m/l went off 3-1 )

Pretty interesting stuff that I have decided to forge on with and pay close attention too. So far it has been valuable.

Croghan won with a 46-1 the other night off a indifferent looking qualifier, he was obviously a bet.

Same note, I bet a 2nd time Varty Thurs. night at the Big M at 6-1 and he backed up through the field.

Any thoughts? And any place that tracks this crap? An excel spreadsheet and stacks of past performances make it difficult. Though I almost hope you say no because if it is out there somewhere, my edge diminishes. I know you can get first off claim and first off lay off stats from trackmaster.


Pull the Pocket said...

Good job Phil.

I used to do quite a bit of that stuff. Off claim numbers way back when etc.

The first good angle in harness-trackers for modern times was the Dbarn stats. Lou, who wrote the B track post before used to do that, and so did I. You would have some 15% trainers who would hit at 3 or 4% in the D. It was a good angle.

It is something I lament about in this business. At betfair you might get every betting stat known to man from them. Then they have their API so enterprising folks use the data to make a product. That product is sold, or given away, and people use it to bet.

If we had a harness website with all past performance data, anyone could pop it into excel, do a bit of minor programming and poof! Betting angles galore.

I did it in excel awhile back. What I had to do was cut and paste all past performances from Standardbred canada, then format it, and mess around. It was like pulling teeth. I got some good stuff, but it was brutal to do. You could spend thousands of hours.

This biz does a lot to make sure it is difficult for us to bet more. Big rakes, no data, hidden video streaming, tracks not accepting other tracks on the internet. On and on. Sometimes I wonder how handles are what they are. They should be even lower!

I think the best way is to track your angle that you see happen, ongoing, by hand. I had one from a trainer, which was second start off a 30 day to 45 day layoff. It was awesome. I hit a 46-1, a few 20-1's and several 8-1's on that angle. I just learned it because I saw this 8th by 22 horse win, and dug deeper. I tracked and bet the rest.

Anyway, long story short. The business does us who want to work at it no favours; and it takes a crapload of work to get all these angles into a playable form.

Im constantly searching for things to help. I am chatting with a couple of excel programmers too, who do t-breds. I would love to be able to do something for harness, because like you mentioned, trainer angles and finding when a horse is on 'go' is a great thing to have. Maybe sometime soon Ill have a plan.

Anonymous said...

That's the beauty of handicapping. Sometimes the best angle to to study the situtations where high percentage angles don't come in and wait for that.

Most Trafficked, Last 12 Months


Carryovers Provide Big Reach and an Immediate Return

Sinking marketing money directly into the horseplayer by seeding pools is effective, in both theory and practice In Ontario and elsewher...