2018 Derby M/L Stab, Last Attempt

[Fairmount1]

My final attempt, purposefully completed pre-draw as in years past, at the Derby off odds when they leave the gate.  A few notes below and a short rant.

1   Justify      2.90
2   Audible      5.90
3   Mendelssohn      7.30
4   Good Magic      9.50
5   Magnum Moon      10.10
6   Bolt D\'Oro      12.30
7   Vino Rossi      15.90
8   Hofburg      21.70
9   My Boy Jack      22.80
10   Enticed      26.40
11   Noble Indy      27.10
12   Solomini      28.80
13   Free Drop Billy      29.30
14   Flameaway      36.00
15   Instilled Regard      37.40
16   Combatant      43.40
17   Promises Fulfilled      45.50
18   Bravazo      51.60
19   Lone Sailor      57.80
20   Firenze Fire      67.90

A very difficult field to assess how the public will wager this one.  Post could seriously affect these odds.  The biggest question is how high or low does Justify go.  I think with all the buzz, undefeated, Baffert trainee, comparisons by him a few times to Arrogate and Pharoah and he will be below 7-2 and hopefully in the 5-2 range as I believe.  The lower his odds pushes all the rest higher.  The next question is how the next 4 are shuffled around and I\'ve seen it all over the place but finally settled as I did above.  

Short Rant:  I\'m no math genius and I\'m not employed in the industry.  But the nonsensical, bizaare and inaccurate things I\'ve read online and on twitter about the m/l by guys and gals paid certainly excellent salaries in the industry is almost umfathomable.  But then I realize, Oh, this is horse racing.  The above line is a correctly to the tenth \"balanced\" line based on the win takeout.  It is all fairly simple math.  But some of these guys even have articles discussing the number of horses as part of the equation......WRONG!  These guys say \"Oh you can fudge it up to 128 or 129 points in the Derby.\"  WRONG.  These guys say that \"oh my line is a little off for the \'sticklers\' out there\"....TRUE!  Why not get it accurate???  It is simple math.  One doesn\'t have to stick to round numbers like the archaic industry does to make a m/l.  There is no rule book that says you only go from 15-1 to 20-1 to 30-1 to 50-1.  It is simple math.  They invented Excel decades ago now I think.  And with the right equations for takeout (to determine the correct points), the right equation for determining points, and a little bit of time and it can be a correctly calculated line even if the guess on the odds of each horse is wrong at least the points (as they call them) will add up correctly.  The reason this is such a big deal is very simple.....If you take any win pool from any race and reverse engineer the math one time with the correct takeout, you will see that it works out exactly to the dollar with the odds.  By \"fudging\" a little you are misleading the public concerning the potential odds on a race.  Of course, I guess a little \"fudging\" when few of these guys are competent is no big deal.  No worries, the racing industry will keep paying them well to do a half-hearted job.  I only do this for one race a year but at least I attempt to make it accurate.  End of rant.  

Good luck.  And tear it to shreds as I say every year!!  I\'ll post next year a comparison to the actual off odds for some accountability.

[Molesap]

Fairmount - what did you use in terms of points for the line - 118? I thought the WPS takeout at CD was 17.5%. I heard someone yesterday say that the Derby line should be 135 points at least to take into account the large field - I never understood why you added the number of horses to make the line.

[Fairmount1]

A long private message sent to you with full explanation.  Phil23 is who showed me the proper way to accomplish this accurately.

It is simple math but neither 117.5 nor 118 and certainly not 135 are the correct number of points.

[Molesap]

Cool - I always wanted to know how to do a line properly.

[Molesap]

I agree with Fairmount on many things and thank him for his efforts with the line. I do think this is tough as it is like a mobile â€" tweak on side and that has an effect on the others. I know it take some work to make an accurate line and he has a great historical knowledge in terms of the betting. As I was fooling around with his line a bit, every time I said oh, this one will get bet more, that correspondingly raises the odds on other entrants

In terms of betting, I think there are four distinct tiers of the odds â€" Group 1 under 15/1, Group 2 between 20/1 and 30/1, Group 3 between 30/1 and 45/1 and Group 4 above 45/1. I was the one who was vocal about Justify being 2/1 on Derby Day, but after playing around with the numbers, I think Fairmount has it right â€" closer to 3/1 than 2/1. I also did not put in the time to make a real line like Fairmount â€" I just estimated the odds and have no idea what the points add up to, but I suspect it is close to what odds makers will do, but it is not the correct way to make a line and account for the points. I have listed them in the order or what I think will be in terms of odds as I think Justify will be the favorite and Firenze Fire will be the longest price on the board. I doubt anyone goes off much over 60/1 though.

I have not heard as much “chatter” on Justify as I thought, but the one horse that seems to be favorably on everyone’s lips in Mendelssohn. He will get bet pretty hard I think. In addition, here is a link to an article about the Derby morning line.

Kentucky Derby Morning Line Musings from Battaglia

1. Justify 3
2. Mendelssohn 6
3. Magnum Moon 8
4. Audible 8
5. Bolt D\'Oro 10
6. Good Magic 12
7. Vino Rossi 15

8. Hofburg 20
9. Enticed 20
10. My Boy Jack 20

11. Noble Indy 30
12. Solomini 30
13. Free Drop Billy 30
14. Flameaway 30

15. Combatant 50
16. Lone Sailor 50
17. Instilled Regard 50
18. Promises Fulfilled 50
19. Bravazo 50
20. Firenze Fire 50

[Mathcapper]

Fairmount1 Wrote:
-------------------------------------------------------
>
> Short Rant:  I\'m no math genius and I\'m not
> employed in the industry.  But the nonsensical,
> bizaare and inaccurate things I\'ve read online and
> on twitter about the m/l by guys and gals paid
> certainly excellent salaries in the industry is
> almost umfathomable.  But then I realize, Oh, this
> is horse racing.  The above line is a correctly to
> the tenth \"balanced\" line based on the win
> takeout.  It is all fairly simple math.  But some
> of these guys even have articles discussing the
> number of horses as part of the
> equation......WRONG!  These guys say \"Oh you can
> fudge it up to 128 or 129 points in the Derby.\"
> WRONG.  These guys say that \"oh my line is a
> little off for the \'sticklers\' out there\"....TRUE!
>  Why not get it accurate???  It is simple math.
> One doesn\'t have to stick to round numbers like
> the archaic industry does to make a m/l.  There is
> no rule book that says you only go from 15-1 to
> 20-1 to 30-1 to 50-1.  It is simple math.  They
> invented Excel decades ago now I think.  And with
> the right equations for takeout (to determine the
> correct points), the right equation for
> determining points, and a little bit of time and
> it can be a correctly calculated line even if the
> guess on the odds of each horse is wrong at least
> the points (as they call them) will add up
> correctly.  The reason this is such a big deal is
> very simple.....If you take any win pool from any
> race and reverse engineer the math one time with
> the correct takeout, you will see that it works
> out exactly to the dollar with the odds.  By
> \"fudging\" a little you are misleading the public
> concerning the potential odds on a race.  Of
> course, I guess a little \"fudging\" when few of
> these guys are competent is no big deal.  No
> worries, the racing industry will keep paying them
> well to do a half-hearted job.  I only do this for
> one race a year but at least I attempt to make it
> accurate.  End of rant.

If you think that\'s bad check out the offshore line Molesap posted last week: Offshore Line as of 4/24/18

The first seven horses already add up to a 125% line. I guess bookies will set the line at whatever they think they can get away with, and they say there\'s a sucker born every minute so I guess they\'re counting on the typical horseplayer not taking the time to run the numbers and realize he\'s getting fleeced.


Molesap Wrote:
-------------------------------------------------------
> Fairmount - what did you use in terms of points
> for the line - 118? I thought the WPS takeout at
> CD was 17.5%. I heard someone yesterday say that
> the Derby line should be 135 points at least to
> take into account the large field - I never
> understood why you added the number of horses to
> make the line.

Fairmount1 Wrote:
-------------------------------------------------------
> A long private message sent to you with full
> explanation.  Phil23 is who showed me the proper
> way to accomplish this accurately.
> It is simple math but neither 117.5 nor 118 and
> certainly not 135 are the correct number of
> points.

It is simple math but there\'s one thing I\'m not sure most linemakers take into account -- the effect of breakage, for which I add ~1.25%.

For CD, after adjusting for their 17.5% takeout and adding 1.25% for the effect of breakage, that works out to:

1/(1 - 0.1750 - 0.0125) = 123% line.

The only reason you\'d want to make a line any higher than that is if you want to leave room for expected scratches, for instance when there are AE\'s that are not expected to get in or when you\'re looking at a futures pool like the Derby that has more horses than will get in. In such cases, until the field is set, the best you can do is to predict which horses are most likely to run and then set the line based on that expected field, with the \"AE\'s\" adding to the line at odds equivalent to the odds of the expected horses they would replace if they were to draw in.

Be curious if Phil23 had any additional insight, but that\'s the way I do it.

[Molesap]

I find this math stuff a fascinating aspect - thanks for the input everyone. I thought it might be interesting to go back and look at the updated offshore line - it is probably even worse now as it appears to me that almost every horse will likely be a better price on Derby Day. To cash the offered 200/1 on Dream baby Dream there has to be a combination of at least seven horses scratching or not entering AND then he has to win the race - that might be the biggest underlay on the board. They have five horses under 5/1 and eight horses at 10/1 or less. I also posted the current odds from the Wynn - very different in some areas. Just think these differences are so interesting...

Offshore
Justify 3/1
Bolt D\'oro 18/5
Mendelssohn 9/2
My Boy Jack 9/2
Magnum Moon 19/4
Audible 13/2
Good Magic 15/2
Hofburg 10/1
Vino Rosso 16/1
Enticed 20/1
Solomini 25/1
Noble Indy 28/1
Flameaway 30/1
Instilled Regard 35/1
Free Drop Billy 40/1
Bravazo 50/1
Combatant 50/1
Lone Sailor 50/1
Promises Fulfilled 50/1
Firenze Fire 75/1
Givemeaminit 75/1
Blended Citizen 100/1
Reride 100/1
Snapper Sinclair 100/1
Sporting Chance 100/1
Restoring Hope 150/1
Dream Baby Dream 200/1

Wynn
JUSTIFY 3
GOOD MAGIC 5
MAGNUM MOON 6
MENDELSSOHN 6
BOLT D\'ORO 8
AUDIBLE 9
VINO ROSSO 14
ENTICED 20
MY BOY JACK 25
NOBLE INDY 28
FLAME AWAY 30
SOLOMINI 30
FREE DROP BILLY 35
HOFBURG 35
INSTILLED REGARD 35
COMBATANT 40
LONE SAILOR 50
RESTORING HOPE 50
BLENDED CITIZEN 75
BRAVAZO 75
FIRENZE FIRE 75
PROMISES FULFILLED 75
DREAM BABY DREAM 250
SPORTING CHANCE 500

[Mathcapper]

Btw, I only indirectly implied it but when dealing with a future\'s pool there\'s of course a time element involved for which the horseplayer should get compensated since in the event the horse doesn\'t make it to the race for whatever reason he forfeits the bet, but as you get closer and closer to race day the line should be a closer and closer reflection of the race-day odds.

The Wynn line seems to be properly taking that into account. That offshore line is ridiculous.

[phil23]

Fairmont - I love this rant - amen.  Couldn\'t agree more.

your rundown is superb - find these sort of discussions so helpful in really digging into the race - fwiw, have it: J-3, M-6, GM-7, A-8, Bolt-9, Moon-10, Vino 12, Hof 17. Which leads to having some really big prices on the stragglers which has not been happening lately so I could easily be wrong. My two thoughts are: A-that Moon will drift and not be single digits-so was pleased to see you have him a bit higher as well - what with almost everyone noticing the bore out last. And B-that Vino gets bet. He was bet hard in both Tampas and now off a win with those two deep pocketed owners kinda think he will be the \"surprise\" bet down on the board at first flash.

Regardless, great discussion and awfully fast, tough race.

Mathcapper - that\'s exactly the way I do it, minus the breakage, which knew was there but had not had a good way/number to account for it - thank you - going to start adding that in each time going forward using your 1.25.

[Fairmount1]

phil23 and Mathcapper,

Is breakage this significant?  When I worked the math backwards way back when on many races I didn\'t recall a significant difference.  In fact, I thought it was right on if memory serves.  Plus, when I create my m/l I actually come up with numbers such as 15.95-1 on Vino Rossi (assigning 5.90 points) and I round the odds down to 15.90 for the sake of the same rounding the track will provide unless at NYRA.  Is there more breakage than me rounding down my odds based on the points I assign than rounding down to the zero?  I know you are far sharper than I at this re: breakage amounts but just wanted to hear more details.  

As for M, GM, A, Moon, and Bolt, those 5 (not 4 as I mistakenly said in my original post) are a bit of a mystery how they shake out.  I actually originally was very strong on Mendy at 5-1 2nd choice although not knowing if his 106 Beyer is in the DRF PP\'s, unknown post currently, and his appearance/buzz the next few days led me to back off.  I went with Audible 2nd choice off the Fla Derby win (Nyquist, Orb, Always Dreaming), JJ jumping ship from the Bolt, the pace setup, Pletcher repeat fresh in people\'s minds, etc.  Magic Moon with Saez, BO, less than 100 Beyer\'s despite being undefeated and the mystery of multiple TAP entries with JJ and JV elsewhere convinces me he is not 2nd or 3rd choice.  But it is tough to know how those 5 shake out exactly.  I\'ll also say I had Hofburg as low as 18-1 in many scenarios and that\'s kind of where I wanted to put him but the lack of huge odds on the bottom of the rung horses likely and also so many in that top 7 so clearly going to take serious dough I couldn\'t put him that low.  But it wouldn\'t surprise me as I\'ve seen him as the wiseguy horse since Gronk caught a bad hamstring apparently, was penalized 15 yards for a late hit, and is sitting this one out.

Good luck!!

[Mathcapper]

Fairmount1 Wrote:
-------------------------------------------------------
> phil23 and Mathcapper,
>
> Is breakage this significant?  When I worked the
> math backwards way back when on many races I
> didn\'t recall a significant difference.  In fact,
> I thought it was right on if memory serves.  Plus,
> when I create my m/l I actually come up with
> numbers such as 15.95-1 on Vino Rossi (assigning
> 5.90 points) and I round the odds down to 15.90
> for the sake of the same rounding the track will
> provide unless at NYRA.  Is there more breakage
> than me rounding down my odds based on the points
> I assign than rounding down to the zero?  I know
> you are far sharper than I at this re: breakage
> amounts but just wanted to hear more details.  


The effect of breakage varies from race to race, depending on the particular odds of each horse and the number of entries in the race. As a result, it\'s not a simple mathematical calculation but rather something that has to be derived empirically, which I did by looking at the final listed odds for hundreds of race charts.

For tracks with $.20 breakage, the average effect of breakage works out to around 1.25%. For tracks like NYRA with $.10 breakage, it\'s around 0.7%.

If you\'re not taking breakage into account when creating a line, it\'s really not a big deal, your line will just sum to a point or two below what it otherwise should be.

[Fairmount1]

As always, much appreciated b/c of all the time and effort you have put into this type of information.  

I will add this to my m/l in all future efforts.  Thank you.  

Hoping you make it on Travers weekend again since you are closer than I am.  I already have the plane tickets and room all set for that weekend for Fri and Sat.

[phil23]

Fair - yep same here as far as those 5 re final odds - really is tough to know exactly where the public will come down. Good point on that Beyer for Mendy too - not sure if it\'s there in print. I\'m swayed by the thorographically pristine pattern of a certain colt but hard to know if John Q P will agree.

Thanks for the more detailed breakdown on the increments for breakage MC, will use that going forward. Ironically the extra juice from the breakage does actually help to make most of these appalling ML\'s ever so slightly \"less\" incorrect (slightly...).

[Fairmount1]

In the comment line for Rayya on Friday it has the comment for the trip as it always does.  Then after that it says Beyer: XX.  (I didn\'t list it obviously but the number is there).  I\'d say it is safe to assume you will see the same for Mendy.  Good luck to you!

[Mathcapper]

Fairmount1 Wrote:
-------------------------------------------------------
> Hoping you make it on Travers weekend again since
> you are closer than I am.  I already have the
> plane tickets and room all set for that weekend
> for Fri and Sat.

Nice, looking forward -- don\'t think I\'ve missed one since Playfellow. Back then it was just a short jaunt from my hometown through the backroads of Hoosick Falls and Schuylerville. A little further now, but fortunately still close enough for even a day trip if I feel the calling..