Below is a little EV calc comparing raising versus calling with a plot showing how our EV changes with the frequency that SB and LMP are willing to overcall.
EVr = b1 * (1bb) +b2 * (2bb) +b3 * (f1*(3bb) - (1 - f1)*(3bb))
+(a1 + b1)*(2bb)
EVc = f2 * (1bb) - (1 - f2)*(1bb) + (a2 + c2)*(1bb)
variables:
a1 = fraction of sb overcalling two cold
a2 = fraction of sb overcalling one bet
b1 = fraction of bb folding to river raise
b2 = fraction of bb calling our river raise
b3 = fraction of bb 3-betting
c1 = fraction of LMP overcalling two cold
c2 = fraction of LMP overcalling one bet
f1 = fraction of the time we are winning when bb 3-bets
f2 = fraction of the time we are beating bb when overcalling
Assumptions:
-we are winning 100% of the time against sb and LMP
-we are winning 100% of the time BB just calls our river raise
For a initial interation I made some assumption as to what some of these parameters should be:
b1 = 10%
b2 = 75%
b3 = 15%
f1 = 10%
f2 = 90%
(will make the independent variable the frequency at which LMP coldcalls)
a1 = 0.25 * c1
a2 = 0.5 * c2
For the graph I assumed that LMP or sb are fairly unlikely to overcall two cold and thus only extend that range from 0 to 0.2. OTOH, they are much more likely to overcall for one bet and thus we have a wider range on the x-axis. Since it is going to be less likely that LMP/SB overcall two cold lets set an semi arbitrary EV for the raising case at 1.4 (seen on the graph by the solid line) to get a feeling for the minimum frequency we need to pick up overcalls to make calling more correct in this situation.
Given all of these assumptions if get and overcall from LMP 40% of the time AND sb 20% then we are essentially indifferent to raising or calling. I would guess that given both of LMP and sb cold called on the flop and didn't raise when the diamonds got there on the turn that they are going to be more likely to overcall this river because they are holding top pairish type hands. Thus I think going for overcalls is going to be more profitable here.