Thanks for the stat Buddybear!

Dogpounder · 01-07-07, 12:01 PM

OK buddybear you're off the hook for this play, but for my legion of fans out there I'll let you know I'm making the following middle play in the Jets/Pats game

NE - 9 1/2 laying $5.00 to make $5.00 per contract
Jets +14 1/2 laying $6.65 to make $3.35 per contract

$11.65 invested per contract pair

$8.35/$1.65 = 5-1 odds

Ganchrow · 01-07-07, 01:38 PM

Originally posted by Dogpounder

OK buddybear you're off the hook for this play, but for my legion of fans out there I'll let you know I'm making the following middle play in the Jets/Pats game

NE - 9 1/2 laying $5.00 to make $5.00 per contract
Jets +14 1/2 laying $6.65 to make $3.35 per contract

$11.65 invested per contract pair

$8.35/$1.65 = 5-1 odds

Based on Historical NFL ATS Push Frequencies this, unlike your first middle, is a negative expected value proposition.

From the historical data we can estimate that your middle will win with probability ~ 15.36% and lose with probability ~ 84.64%. Since you're receiving $8.35 / $1.35 on your money this corresponds to an expected loss of about 6.91% of your net at-risk amount. In other words for each pair of $10 notional contracts purchased, you expect to lose ~ $0.114.

Dogpounder · 01-07-07, 04:28 PM

Hey thanks for your input, I'm glad to now know those stats.

Of course I wouldn't make such a play just randomly, I do it usually when a game feels like it will play out in a certain range.

Sometimes they slide into the range at the end of the game, sometimes like this game an interception returned for TD slides it out of range.

No problem, I'm more concerned about being correct about the point range.

Give me 10 shots at that situation, (Pats up 14, 5 minutes left, Jets ball at own 30) and I like my chances to hit the middle 3 times at least, good profits when I'm right about handicapping the range.

BuddyBear · 01-07-07, 09:06 PM

No problem...not exactly sure what I did though....

Ganchrow · 01-07-07, 09:46 PM

Originally posted by Dogpounder

Give me 10 shots at that situation, (Pats up 14, 5 minutes left, Jets ball at own 30) and I like my chances to hit the middle 3 times at least, good profits when I'm right about handicapping the range.

I heartily agree that this full-game middle would have be an excellent bet had you been able place it with 5 minutes remaining in the game.

However, since the bet was ostensibly made before the game began, then if the historical record was to be believed (and possibly it shouldn't have been) then this particular bet was an expected loser.

pags11 · 01-08-07, 02:28 AM

buddybear is a skilled capper...one of the better number crunching handicappers I've ever been around...

Dogpounder · 01-08-07, 04:47 AM

Buddybear, trust me - you da man!

Mr. Ganchrow, once again I appreciate your interest in what I am writing, even if I am not getting your wholehearted support.

Let's take the number NE - 9 1/2 and consider 4 game scenarios.

1) Jets are better or luckier and win SU
2) Jets play tough but fall short. They lose the game but cover 9 1/2
3) NE doesn't break much of a sweat and wins comfortably (10-17 points)
4) NE woke up angry as hell that morning and had the ability and desire to crush the poor Jets (18 + points)

I hope you don't disagree that one would be more to likely lay the points if one's analysis (before the game starts) indicates a much lower probabilty of scenario 1 occurring than scenario 4 occuring. Likewise, if one thinks scenario 1 is considerably more likely than scenario 4, one would be more inclined to take the points.

If you don't agree up to this point, you can ignore the rest of this.

Now in some games, one's analysis (before the game starts) can lead to BOTH scenarios 1 and 4 having lower probabilities of occurring then a typical average random game.

In these cases, scenarios 2 and 3 become more likely, i.e. the probablity of the final outcome getting closer to the spread number gets greater (if your handicapping is correct). Time to play a middle, IMHO.

It sounds like in your handicapping world, the only time to make a play is when your analysis indicates a team is stronger than the spread number indicates, This is the only time you are able to see value, if a line is - 9 1/2 when your analysis indicates -11 is the correct number.

I am saying there is more than one way to skin a cat, and this is another way to get an overlay. If the probabilities of scenarios 1 and 4 are reduced, then the probabilities of scenarios 2 and 3 increase, making a middle (at the right price, which I believe I was getting) a considerable overlay.

But it looks like we'll just have to agree to disagree on this one.

Ganchrow · 01-08-07, 07:04 AM

Originally posted by Dogpounder

Mr. Ganchrow, once again I appreciate your interest in what I am writing, even if I am not getting your wholehearted support.
-snip-

But it looks like we'll just have to agree to disagree on this one.

This is a subject which is in fact quite dear to me -- specifically the degree to which deviations in outcome volatility impact the value of half-points. This is particularly interesting because to my knowledge no sportsbooks factor outcome volatility into their half-point valuations.

I'd point out that it's important to recognize that through this analysis, one is not actively determining a MOV expectation (despite the point of view you mistakenly attributed to me in your third paragraph to last), but rather is assuming a priori a specified MOV expectation (in this case 9½).

So if I'm accurately assessing your understanding of this point correctly (and it is certainly a bit of a fine point, rather prone to cargo-cultish interpretations) then I'd say you did have a positive expectancy bet, insofar as your estimate of the outcome volatility was sufficiently accurate. As it relates to this post, the explanation you appear to have given would certainly fall under the aegis of "times the historical record is not to be believed".

IMHO, the best way to approach a quantitative topic such as this is probably to start off with a strong knowledge of the basics -- in this case an intuitive understanding of the determination of half-point valuations assuming constant outcome volatility. Once you're totally comfortable with this methodology, is when you can really start properly applying ANOVA to valuations. That of course is when the topic starts getting really interesting.

Anyway though, good job and I do apologize if the salience of your argument was swept up by my more forum-friendly simplification of the issue.