[lordswing]

Lakers/6ers are 5.5 -101/-109 at Pinny
6ers +6 -121, +6.5 -131

ShitBook #1 has +6 -102, giving me 5.24% Edge, 50.93% WP
ShitBook #2 has +6.5 -110, giving me 5.17% Edge, 55.09% WP

Kelly based on WP
Mr. HalfKelly is suggesting .44x for +6 -102 and 2.85x for +6.5 -110

Ideally if I wasn't tapped out at ShitBook #2, I would just hit up +6.5 -110 and kiss 2.85x away as the Lakers run away with it. But I am tapped out, and see I can buy a half point for 10 cents (+6.5 -112) at ShitBook #1, increasing that .44x to 2.44x.

Kelly based on Edge
Sixers +6 -102 Edge: 5.24%, 2.67x
Sixers +6.5 -112 Edge: 4.28%, 2.40x
Sixers +6.5 -110 Edge: 5.17%, 2.85x

I'm just confused as to how Mr. HalfKelly has to 2 different suggestions on the same line (Sixers +6 -102, .44x (50.93% WP) vs. 2.67x (5.24% Edge)), but is consistent on +6.5 -110

e: Just noticed I was using the HPC with Pinny's 5.5 lines, going to redo wrt 6 and 6.5 I guess.

Pinny's line 6 108/-121 (Edge vs. WP)
6 -102 5.45%, 51.03% (2.78x vs. .54x)
6.5 -110 5.36%, 55.19% (2.95x vs. 2.95x)
6.5 -112 4.47%, 55.19% (2.51x vs. 2.50x)

Pinny's line 6.5 118/-131 (Edge vs. WP)
6.5 -110 5.54%, 55.28% (3.05x vs. 3.05x)
6.5 -112 4.64%, 55.28% (2.60x vs. 2.60x)

[Data]

If you have to use HPC you should NOT be using Kelly.

[Tomato]

If you have to use HPC you should NOT be using Kelly.

Why

[FreeFall]

If you have to use HPC you should NOT be using Kelly.

yeah I'm not sure where you are going with this one.

[lordswing]

I'm going to guess because I have to rely on the HPC to get my "edge" as well as the HPC not being 100% accurate

When lines such as 5.5, 6.0, and 6.5 are available at various books - I should find the implied odds based off of Pinny's drop down menus right? i.e. Using Pinny's 6.5 vig to get true odds, then compare ShitBook's lines to that.

Of course, the true answer here would've been to please Lakers -12 50x

[goblue12]

I'm going to guess because I have to rely on the HPC to get my "edge" as well as the HPC not being 100% accurate

When lines such as 5.5, 6.0, and 6.5 are available at various books - I should find the implied odds based off of Pinny's drop down menus right? i.e. Using Pinny's 6.5 vig to get true odds, then compare ShitBook's lines to that.

Of course, the true answer here would've been to please Lakers -12 50x

Yes.

HPC hasn't updated push frequencies in a year or two.

[Data]

I'm going to guess because I have to rely on the HPC to get my "edge" as well as the HPC not being 100% accurate

This.

In NBA, the half-point values considerably differ for home and away teams. The HPC does not account for this and therefore guarantying every number to be somewhat erroneous making it unusable for cases where one needs a fairly high precision while estimating small edges.

[IrishTim]

Data, what do you think the best way to calculate the value of half points? I personally am not a big fan of the small landing area approach (even with a radius to broaden the sample size).

[Data]

Data, what do you think the best way to calculate the value of half points?

One way or another, and I am not saying it is easy, one should come up with a distribution that takes estimated team scores as parameters and calculates the probabilities for every reasonable outcome. So, instead of calculating half point values the bettor will have the ability to calculate his own odds for a range of lines.

To illustrate, the model comes up with estimated scores of 93.18 and 86.81 for teams A and B in a given matchup. We feed these numbers into Big Mean Calculator Powered by the Incredible Beauty of our Distribution Function and that "calculator" gives us the odds for a range of spreads from, say, 3 to 10 with a half point steps. Now, we going to take the available lines from the books (say, +6 -102 and +6.5 -110) and compare the with our odds. So, for every spread we can see what is our edge. Notice, that now we can see how much every half point is worth just by looking at the lines "The Calculator" gave us. So, while the value of those half points is totally useless for us at this point, this, I think, is the best way to calculate them.

[Tomato]

Of course, the true answer here would've been to please Lakers -12 50x

Welcome to basketball. You won't enjoy your stay.

[horsiehung]

lets not forget about the team winking to each other, letting a -13 or more dog hang around, only to be beaten by single digits...lakers last week