ParlaySize in calcKelly()

[deltap]

Ganchrow,
I am trying to rewrite in vba your Kelly calculator, but I can not understand what you do with the binary numbers in ParlaySize. Can you please explain this?
Thanks

[Ganchrow]

Ganchrow,
I am trying to rewrite in vba your Kelly calculator, but I can not understand what you do with the binary numbers in ParlaySize. Can you please explain this?
Thanks

I assumme you're referring to the ParlaySize function :

function ParlaySize(lParlay) {
	var sParlay = dec2bin(lParlay,0);
	var lSize = 0;
	for (var i = 0; i < sParlay.length; i++) {
		lSize += parseInt(sParlay.charAt(i));
	}
	return lSize;
}

Where the function dec2bin(x,m) converts the base-10 value x to a binary string of minimum length m.

The idea is that a given parlay number (of type long) when converted to binary corresponds to a unique parlay where the '1' digits refer to active games within the parlay.

So for example for 4 independent events:

• Parlay #7 (binary 0111) would refer to the 3-team parlay of game #s 1,2, & 3 (counting right-to-left).
• Parlay #9 (binary 1001) would refer to the 2-team parlay of game #s 1 & 4
• Parlay #4 (binary 0100) would refer to the 1-team parlay (i.e., the single) only of game # 3.

All the ParlaySize function does is loop through the digits of of the binary string, count the number of "1"'s, and then return that value.

[deltap]

Thanks Ganchrow, I will send you the code when finished.
Thanks again

[deltap]

Kelly calculator in VBA for Excel

Hi Ganchrow,
Attached is an excel file with the VBA code. This is my first attempt and is a bit slow. Any recomendations welcomed.
calcKelly5.xls

[Ganchrow]

Your results don't appear correct for full Kelly.

Do you want to walk us through what you have?

[Ganchrow]

Also, is there any particular reason you prefer the unbounded exact algebraic solution to a constrained and optimized one?

[deltap]

Your results don't appear correct for full Kelly.

Do you want to walk us through what you have?

Ok lets do an example.
lets suppose o1=1.31, o2=1.21, o3=1.23 and p1=0.83, p2=0.83, p3=0.83.
then by substitution to Kelly formula (pi*oi)/(pi-1) (i=1,2,3) we get k1=0.29569, k2=0.03968, k3=0,10869 and k1*k2*k3=0.0000012 which does not equal or even close to your Kelly calculator. Most probably I do something wrong with the Kelly Formula.
What is the Kelly formula I should use?
I even tried the formula from the Javascript of your Kelly calculator but it gives me the same results.

[deltap]

Also, is there any particular reason you prefer the unbounded exact algebraic solution to a constrained and optimized one?

what is "unbounded exact algebraic solution" and what is "a constrained and optimized"

[Ganchrow]

Ok lets do an example.
lets suppose o1=1.31, o2=1.21, o3=1.23 and p1=0.83, p2=0.83, p3=0.83.
then by substitution to Kelly formula (pi*oi)/(pi-1) (i=1,2,3) we get k1=0.29569, k2=0.03968, k3=0,10869 and k1*k2*k3=0.0000012 which does not equal or even close to your Kelly calculator. Most probably I do something wrong with the Kelly Formula.
What is the Kelly formula I should use?
I even tried the formula from the Javascript of your Kelly calculator but it gives me the same results.

o₁ = 1.31
o₂ = 1.21
o₃ = 1.23
p₁ = p₂ = p₃ = 0.83

The single bet full-Kelly formula is:

k_i = (p_i*o_i-1)/(o_i-1)

Hence:

k₁ = (0.83*1.31-1)/(1.31-1) ≈ 28.161%
k₂ ≈ 2.048%
k₃ ≈ 9.087%

So k₁ * k₃ * k₃ ≈ 0.0524%

etc.

[deltap]

Thanks Ganchow

[deltap]

Thanks again. This is the new file.

[Ganchrow]

Thanks again. This is the new file.

Looks good.