Overlays

[Magister Ludi]

Overlay – A situation in which offered odds are greater than expected odds. Please see the following race which recently ran:

https://www.equibase.com/static/chart/pdf/AQU031013USA.pdf

Assume that the final odds were available to you at post time but that you didn’t know what the outcome of the race was going to be. Which horse(s) do you think were overlays? (Please see poll above). I voted for #2 and #6.

[Red Knave]

AQU01 (Main.AQU01) Bet Int=7
sort:Pct+$Net Time: 17:47:57 Min: 4:57:57+
Prime Cont=4 FH Cont=50% +1
Prg Horse Con Odds Prob $Net Opt% Gr |
2 BIRD FU C 14.20 46.3 $14.06 42.46% A+ |
7 LUCY STRAGMORE C 3.65 25.0 $2.32 4.38% A |
4 PROUD SIREN C 6.30 11.0 $1.60 -3.17% B- |
3 PRECIOUS FRANCA f 2.75 9.1 $0.68 -24.00% F |
5 E STREET BOURBON 1.35 7.6 $0.36 -60.74% F |
8 TINA'S NOTE 22.20 1.0 $0.48 -3.42% F |
6 MY BROWN EYED BABY 22.10 0.0 $0.01 -4.50% F |


2 and 7 for me

[Stillriledup]

AQU01 (Main.AQU01) Bet Int=7
sort:Pct+$Net Time: 17:47:57 Min: 4:57:57+
Prime Cont=4 FH Cont=50% +1
Prg Horse Con Odds Prob $Net Opt% Gr |
2 BIRD FU C 14.20 46.3 $14.06 42.46% A+ |
7 LUCY STRAGMORE C 3.65 25.0 $2.32 4.38% A |
4 PROUD SIREN C 6.30 11.0 $1.60 -3.17% B- |
3 PRECIOUS FRANCA f 2.75 9.1 $0.68 -24.00% F |
5 E STREET BOURBON 1.35 7.6 $0.36 -60.74% F |
8 TINA'S NOTE 22.20 1.0 $0.48 -3.42% F |
6 MY BROWN EYED BABY 22.10 0.0 $0.01 -4.50% F |


2 and 7 for me

I thought the favorite was the biggest underlay on the board. She raced poorly in her previous start against what was probably a weaker field, the public bet on the trainer change and hoped she would be moved up a few lengths but she wasnt going to be able to win unless the trainer moved her up dramatically overnight.

I would say the 2nd place finisher and 4th place finisher were the biggest overlays.

[Robert Goren]

With the :5: horse which I felt had no chance taking so much money, they all were. The trouble was I was not in love any of the others either and end up on the:7: . I find it really tough to pass a race when I think the favorite is a dog, but this was one those race.
If this race had been at Tampa instead of AQU, then the :5: is a lock. Ness's magic doesn't as well in NY as it did Florida.

[MJC922]

Some output below from a spreadsheet I was experimenting with last month to generate a preliminary line. So far it's only setup to use the last five ratings and the projected early positions. Haven't got around to incorporating the trainer into the mix yet but the trainers for the 3 & 4 we're graded below avg for me so I suspect those very marginal overlays would evaporate anyway. The trainer for the 7 I have as average so no penalty there but it's still a very slim overlay at best. I suspect the 5 would be a solid underlay even after Ness is factored in.

Num MyOdds Odds
2 18-1 14.2
3 2-1 2.75
4 9-2 6.3
5 3-1 1.35
6 72-1 22.1
7 3-1 3.65
8 99-1 22.2

[riskman]

My contenders in order and my odds line:

7- 3.8

3- 4.5

2- 6.0

4- 7.5

Did not bet the race .The 7 a slight overlay--the 2 F/O 14.2 nice value.

[Tom]

The 7 and the 8.
The 7 by such a small amount I wouldn't consider it a real good overlay.
the 8 was just not a contender imho, so I don't care if it went off at 50-1, I would not have bet it. I would use both in a contest with play money, though.

[Rise Over Run]

Although not specified, I'm pretty sure the OP was looking for responses associated with Race #9. This is based on the program # options in the poll.

[Magister Ludi]

Although not specified, I'm pretty sure the OP was looking for responses associated with Race #9. This is based on the program # options in the poll.

Good catch! Now the poll is absolutely useless, but there are still some interesting comments on the [incorrect] race that I linked.

[Pensacola Pete]

Maybe you should post it again, using the single chart link.

[DeltaLover]

Interesting topic. Past performances would have been nice to see as well.

My opinion is that the concept of an overlay is not well understood and usually misused among horse players.

Considering an overlay any horse starting with higher odds than what it 'should' have been, making not adjustments for its favoritism. Based in this approach a horse can be an overlay even if it is the odds one favorite of the race or one of the top choices. I do not agree with this point of view, since an obvious horse that should had been 2-1 is offered by the pools at 4-1 most of the time does not signify an overlay but a dead horse who is not taking the proper action. The opposite is true as well, a horse who 'should' had been a 9-1 shot is offered as a 5-2 most likely is not an underlay but an overlay.

Whenever there is a significant discrepancy between our line and the public's the latter will be proven correct in most situations.

It is impossible to have an objective opinion about the true odds of a horse without considering betting action. This process should be weighted towards the top choices since this is where the crowd in making the biggest mistakes or biggest hits! Based in this it is impossible to conclude that a 30-1 actually represents an overlay if we are not sure that a serious betting anomaly is present on the top betting choices. Races like for example the recent Tampa Bay Derby where the crowd was correct in its main selection have very little to offer since the margin of error on the rest of the field will never be large enough to create opportunities beating the takeout by a significant margin. On the other hand races like the recent Gulfstream Park Handicap where the favorite Fort Larned was matching a factor that is overbet by the public (http://alogatas.wordpress.com/2013/02/10/leo-dont-be-romantic/) was presenting betting opportunities making it correct to bet most of the other starters with not much further handicapping needed (of course I know that Ford Larned lost the rider and based in what we saw he would most likely win the race but this does not change my argument which does not imply that all of these horses are going to loose rather that they will loose to a larger extend than what the crowd thinks...)

As far as the race examined here I can see that two of its starters where both under 3-1 and from them we have to start our handicapping process. If we conclude that they both represent underlays then it is easy to make a profit (assuming we are correct in our assessment). More than this if one the next betting choices (4 and 10) represent overlays based in their profiles, meaning he is not getting the required action we can also eliminate him narrowing even more the list of our contenders. Note though, that the same does not apply to higher odds starters since they are not so visible making more possible for the crowd to make a mistake on them. If this is the case and we can eliminate both favorites : 6 and 1 and one of 4 and 10 we are left with five starters:

3: 8.7
2: 36.5
7: 6.3
8: 40.5
10: 4.10

Knowing that some of them represent overlays. The cumulative value of betting this group of horses is slightly more than even money so if we are correct at least half of the time in our handicapping we will have a break even bet. The betting should go as follows:
10: 40.3 %
7 : 28.2 %
3 : 21.2 %
2 : 5.5 %
8 : 4.9 %

Meaning for example that we need to bet the #10 40.3% of the our total bet. Obviously if we can eliminate one of 10 - 7 - 3 we can increase the return of our bet, although we need to have a good understanding of the risk involved since we now it is more possible for the crowd to underestimate than to overestimate a starter. With this I mean if we were able to eliminate 10 and 3 we would have been left with only:

7: 6.3 73%
2: 36.5 14%
8: 40.5 13%

In this case betting all three starters we have a cumulative return of 4.3 - 1 while by eliminating the two longshots 2 - 8 we are only going to 6.3 - 1. For us to single the 7 we need to believe that the probabilities of any of the 2 or 8 to beat the 7 are less than 27% something that most of the times it will be very hard to justify.

[Red Knave]

Although not specified, I'm pretty sure the OP was looking for responses associated with Race #9. This is based on the program # options in the poll.
Geez, I didn't see that either.

AQU09 (Main.AQU09) Bet Int=8
sort:Pct+$Net Time: 16:23:52 Min: 0:28:07
Prime Cont=4 FH Cont=50% +1
Prg Horse Con Odds Prob $Net Opt% Gr |
6 CHRISTOPHER'S JOY C 2.75 46.3 $3.47 26.73% A+ |
7 WANSTEAD C 6.30 15.3 $2.23 1.83% A |
4 DAN'S GOLD C 4.40 13.7 $1.48 -5.91% C |
3 FLO'S GREATNESS C 8.70 13.6 $2.63 3.62% A+ |
10 LITTLE DALE f 4.10 7.6 $0.78 -14.88% F |
2 DABA DABA 36.50 2.5 $1.84 -0.22% B+ |
1 COPY MY SWAGGER 2.85 0.8 $0.06 -34.04% F |
8 TOP YANK 40.50 0.3 $0.27 -2.14% F |

Now it's 6, 3 and 7

[MJC922]

From my experience, I've found it necessary to have a line on my horses. It doesn't have to be a line on every horse (which would be ideal as it opens up a lot of possibilites for lower percentage but potentially higher ROI action-type plays) but at least I have to have a line on my selection. I agree that a line sometimes has to be adjusted for tote action as tote action itself is a factor in racing along with a hundred others, however adjusting for tote action should be the exception rather than the rule. An example where tote action rises in importance would be when you have a horse making its second lifetime start and the night before you were making this horse even money to win the race. The next day the betting opens and a firster from an excellent barn opens up at 2/5. Well now you're getting 2-1 on your horse and it becomes an 'overlay' prime bet IF you stand firm on this assessment but you can't really ignore the tote action. To be pragmatic about it you have to admit your horse's chances drop maybe from 50% to 40% or less. We know the crowd's line is accurate longterm but that doesn't mean a darn thing in the short term. I have seen marginal horses take thousands of dollars from an offtrack teller desperate to get out for the day because they've been betting out of their drawer. Individuals go overboard every day to various degrees and you have to always be ready to capitalize. A winning player's own line on his/her selections will be accurate longterm as well. When the two lines differ an overlay situation occurs. Let's remember though, when you think the crowd is way off they would love to be betting into your line too, but since both lines are longterm accurate the actual win% will inevitably fall someplace in the middle.