The Problem with Morning Lines

[thaskalos]

Originally Posted by DeltaLover

Thask, in my models (especially the one we talked about yesterday night), I always need a rating that results to a valid ranking (where higher ranked horses win more than lower) and the one that I am currently using for this purposes is BPP. I have to believe that M/L can be used as an alternative, although in this case we have the additional complexity of refining our models in a per racing secretary (or whoever else is making the ML)..

When using the BPP, you are using a method which you know is being meticulously calculated, and consistently applied. But when you use the ML...you aren't even sure how serious its originator is about the task that he is performing...or even if it's the same guy who is always doing it. Is the morning line-maker someone who takes pride in his work...or is he someone who is overworked and underpaid, and breezes through the process so he can go home early to spend some time with the kids?

In view of this...how can the ML be considered a viable alternative to the BPP?

[DeltaLover]

Originally Posted by thaskalos

When using the BPP, you are using a method which you know is being meticulously calculated, and consistently applied. But when you use the ML...you aren't even sure how serious its originator is about the task that he is performing...or even if it's the same guy who is always doing it. Is the morning line-maker someone who takes pride in his work...or is he someone who is overworked and underpaid, and breezes through the process so he can go home early to spend some time with the kids?

In view of this...how can the ML be considered a viable alternative to the BPP?

Your concerns are valid and to some extend I maintain the same views.

Still, I can also detect that ML is behaving similarly to BPP, producing very correlated rankings. Of course there seem to be some differences when it comes to ROI (I have a theory why this might be happening)...

Again, I am not using ML in my models (at least not as a basic entry point) but I consider it possible to develop an approach based on them..

[thaskalos]

Originally Posted by DeltaLover

Your concerns are valid and to some extend I maintain the same views.

Still, I can also detect that ML is behaving similarly to BPP, producing very correlated rankings. Of course there seem to be some differences when it comes to ROI (I have a theory why this might be happening)...

Again, I am not using ML in my models (at least not as a basic entry point) but I consider it possible to develop an approach based on them..

I could be wrong...but I've always thought that the competent horseplayer should have a better opinion than that of the ML. The ML is wrong so often, and so often in need of "correcting", that the horseplayer would be lost if he lacked the skill to recognise these mistakes and rectify them. And if the horseplayer does indeed have a better opinion than the ML...then the ML becomes unnecessary to him.

And I don't buy the argument that the ML still influences the betting in anything but a superficial way. The unsophisticated players have thinned out in this game, and the ML followers were among the first casualties, IMO.

[DeltaLover]

Originally Posted by thaskalos

I could be wrong...but I've always thought that the competent horseplayer should have a better opinion than that of the ML.

This does not mean that he should not consider it.

Every decision we are facing, depends on comparing our subjective opinions against the public and having an opinion about the validity of the ML in not an exception..

[ReplayRandall]

Originally Posted by GameTheory

We've had this discussion before a few times, so I will quote myself from an old post:



That was from this thread: http://www.paceadvantage.com/forum/s...d.php?t=116932

GT, got to give you props for a great detailed post......Fully agree.

[thaskalos]

Originally Posted by DeltaLover

This does not mean that he should not consider it.

Every decision we are facing, depends on comparing our subjective opinions against the public and having an opinion about the validity of the ML in not an exception..

I have considered it...and I have rejected it. That is my own opinion, of course...and I would never presume to impose my opinion on others. All we can do is offer opinions on this site; opinions that are hopefully based on our own playing experiences. Defending an opinion isn't the same as imposing that opinion on someone else.

[ReplayRandall]

Originally Posted by thaskalos

I could be wrong...but I've always thought that the competent horseplayer should have a better opinion than that of the ML. The ML is wrong so often, and so often in need of "correcting", that the horseplayer would be lost if he lacked the skill to recognise these mistakes and rectify them. And if the horseplayer does indeed have a better opinion than the ML...then the ML becomes unnecessary to him.

And I don't buy the argument that the ML still influences the betting in anything but a superficial way. The unsophisticated players have thinned out in this game, and the ML followers were among the first casualties, IMO.

I'm capable, as are others, of making their own line. I never look at the morning line intentionally......ML is outdated and obsolete in this day of high-tech enlightened handicapping/wagering.

[Dave Schwartz]

Has anyone pointed out the way in which the morning line is beneficial to the competent horseplayer?

If you are a guy who looks for value based upon the tote board, these days you are in a world of hurt unless you can predict what the odds are going to be instead of what they are at the time you make your bet.

The exception to that is, of course, if you lay in the weeds looking for horses that you PERCEIVE should be (say) 3/1 and are (say) 8/1 as they load the gate. IOW, you believe the horse has a 25% chance of winning and is going to pay around $18.

This would translate into a $4.50 $net (i.e. +125% advantage).

Of course, since nobody ... NOBODY... is a +125% handicapping your probability estimate would likely be wrong by a near landslide. If we assume that you might be a +20% handicapper, then your hit rate is likely around half the projected 25% on horses that actually go off at 8/1.

Logically, as the odds go down, the hit rate would rise.

The point is that while everyone is disdaining the ML as being a worthless projector, it is actually almost as good as (say) the live tote at around 8 minutes to post. (Anecdotal; I have no evidence to offer but sounds about right to me.)

POINT: Anyone attempting to make a value-based decision well before post time has 3 choices:
1) Build a rigorous model that predicts final odds using captured minute-by-minute tote data (a complicated endeavor)
2) Build a rigorous model that predicts final odds using a collection of factors (a complicated endeavor)
3) Make their own ML (do not confuse this with "fair odds" as people like to call it)
4) Use the published morning line

About #1 & #2: If you have not built a rigorous probability model it is highly unlikely that you are going to take the time to build a final odds model.
(Special note: We're talking about thousands of races here and lots of number crunching.)

About #3: "Make their own morning line" does not mean saying, "Well, I think this horse is going to be 3/1." What makes anyone think that they can do this any better than the guy at the track?

About #4: At least the published ML is a somewhat consistent commodity. As bad as it can be at times, it is still better than anything else.

It also has the added advantage over #3 of being "self-fulfilling" because the bonehead public will collectively TRY to adjust the tote to match the line maker. Understand I am not talking about the sophisticated players as we find here (and I am not being sarcastic).


Now, as with any skill-based endeavor, there will always be the rare example of the handicapping savant that just does this stuff in their head and makes it work.

It is that "makes it work" part - which means "beating the game" - that is difficult.

I do not believe that there are very many such people. If you are one of those people, then I congratulate you and have great respect for you. In such a case, the ML would have absolutely no value for you.

Dave

PS: Just saw this:

Originally Posted by GameTheory
We've had this discussion before a few times, so I will quote myself from an old post:

That was a great post. Said it better than I did.

[Capper Al]

Originally Posted by DeltaLover

I use rankings instead

Ditto

[thaskalos]

Originally Posted by Dave Schwartz

Now, as with any skill-based endeavor, there will always be the rare example of the handicapping savant that just does this stuff in their head and makes it work.

It is that "makes it work" part - which means "beating the game" - that is difficult.

I do not believe that there are very many such people. If you are one of those people, then I congratulate you and have great respect for you. In such a case, the ML would have absolutely no value for you.

Dave

I manage to show a decent profit while still maintaining my disdain for the ML...but I hardly consider myself to be a "handicapping savant". I work hard, and burn the midnight oil, to do what I do. To me, the term "handicapping savant" should be reserved for those who manage to turn a handsome profit at the horses, while still leading a normal lifestyle.

[Dave Schwartz]

...but I hardly consider myself to be a "handicapping savant".

Thaskalos,

I would contend that is precisely the level you function at as the winner you are.

That is why there are not many people doing it.

As for "midnight oil," there are lots of people that burn lots of 55 gallon drums of that stuff and also keep refilling their wagering account.

[thaskalos]

Originally Posted by Dave Schwartz

Thaskalos,

I would contend that is precisely the level you function at as the winner you are.

That is why there are not many people doing it.

As for "midnight oil," there are lots of people that burn lots of 55 gallon drums of that stuff and also keep refilling their wagering account.

Well...the QUALITY of the work counts too, you know. Some painters paint houses...while others paint masterpieces which hang on museum walls.

[TheOracle]

The Morning Line favorites have taken a severe beat down since Preakness Weekend at Belmont.

Since May 15th up until yesterday the Morning Line Favorites had won only 11 out of 39 races.

The Morning Line 2/1's have really been taking a beating winning only 1 of their last 10 attempts since May 15th as of yesterday.



When just prior to that the 2/1's had done their best work winning 8 of 10 attempts from May 6th to May 10th.



It's interesting to note that since May 15th, on the Dirt surface as of yesterday, the Morning Line Favorites have been doing much better at Belmont




This is what happened today on the Dirt at Belmont:




All / 5 ( represents Morning Line Odds ending with 5 i.e. 2/5, 3/5, ,,,etc.)


I would have to say that specifically the Morning Line Favorites, at least at Belmont, win at a rate that is cyclical sometimes they are on fire and other times they are just stinkers.

[Dave Schwartz]

(No surprises there. It is a horrible factor for predicting winners.)

Oops. Misunderstood the tables.


How about a report on the value of predicting the final public 1st, 2nd, 3rd choices over a few thousand races?

Perhaps look at the R2 of the ML rank and the PubCh rank.

[Some_One]

Originally Posted by TheOracle

The Morning Line favorites have taken a severe beat down since Preakness Weekend at Belmont.

Since May 15th up until yesterday the Morning Line Favorites had won only 11 out of 39 races.

The Morning Line 2/1's have really been taking a beating winning only 1 of their last 10 attempts since May 15th as of yesterday.



When just prior to that the 2/1's had done their best work winning 8 of 10 attempts from May 6th to May 10th.



It's interesting to note that since May 15th, on the Dirt surface as of yesterday, the Morning Line Favorites have been doing much better at Belmont




This is what happened today on the Dirt at Belmont:




All / 5 ( represents Morning Line Odds ending with 5 i.e. 2/5, 3/5, ,,,etc.)


I would have to say that specifically the Morning Line Favorites, at least at Belmont, win at a rate that is cyclical sometimes they are on fire and other times they are just stinkers.

Homer : Ah, not a bear in sight. The Bear Patrol must be working like a charm.
Lisa : That's specious reasoning, Dad.
Homer : Thank you, honey.
Lisa : By your logic, I could claim this rock keeps tigers away.
Homer : Oh, how does it work?
Lisa : It doesn't work.
Homer : Uh-huh.
Lisa : It's just a stupid rock.
Homer : Uh-huh.
Lisa : But I don't see any tigers around here, do you?
Homer : Lisa, I want to buy your rock.