| |
BGonline.org Forums
Janowski centered cube equity ?
Posted By: MaX In Response To: Janowski centered cube equity ? (David Rockwell)
Date: Friday, 6 March 2009, at 9:32 a.m.
I think I get what gnubg uses: you compute the live cube centered equity and knowing the dead cube one, you just interpolate between the two acconding to the cube life index x. This implies that the expression for the general cube model centered equity is linear in x. Janowski's one is not linear in x.
What I get is:
Cv*(p*(W+L)-L + (p*(W+L+2)-L-1)*x/3);
In the paragraph "Centere cube equities" he says (quote):
"No simple formula is available to calculate cube-centred equities, but we do know four points where the cubeless winning chances and corresponding equities are known. These are given below, in order of increasing probability (and equity):"
The 4 points/equities are (TakePointA,-1), (InDoublePointB,E1), (InitialDoublePointA,E2) and (CashPointA,+1).
Now E1 and E2 are computed using the expression of the general cube mode equity owning and with unavailable cube, evaluated at IDPB and IDPA, fine. But how do you compute IDPB and IDPA ?
I don't get it: do we define the centered equity and then deduce the initial double point or do we do the other way round ? And if we start from the IDP, how do we compute them ?
Also, it is not true that in the general model with x=0 (i.e. in the dead cube model) the centered equity at the Take/Cash point is -1/+1. This is only true in the live cube model.
MaX.
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.