Special Bug Pages

Showing posts with label EV of shoving all in. Show all posts
Showing posts with label EV of shoving all in. Show all posts

Friday, March 2, 2012

The Quiz Question That Won't Die (Q #48)

Okay, once more into the breach. Here's the math that shows shoving all-in with AKo preflop against the two TAgs in Quiz Question #48 is negative EV.

First, let's look at TAg #1, who was the original opener in MP:
  • TAg #1 Original Raising Range= 15.20% (22+,ATs+,KTs+,QTs+,JTs,ATo+,KJo+,Qjo) 
  • TAg #1's 4bet calling range= 0.90% (KK+) 
  • Probability that TAg #1's will call our shove= 5.9%  (this is just 4bet calling range divided by original opening range)
  • Therefore, probability that TAg #1 will fold to all-in shove= 94.1% (this is just 100% minus the probability that we get called)
Similarly, we can then make some assumptions and calculations about TAg #2:
  • TAg #2 Original ReRaising Range= 5.00% (99+,AQs+,Ako)
  • TAg #2's 4bet calling range= 0.90% (KK+) 
  • Therefore, probability that TAg #2's will call our shove= 18.0%
  • Therefore, probability that TAg #2 will fold to all-in shove= 82.0%
All of this means:
  • Probability that TAg #1 and TAg #2 will both fold = 77.1% (To calculate this, just multiply the two probabilities that the TAgs will fold together; in this case: 82% x 94.1%)
  • Therefore, probability that we get called by at least one TAg =  22.9% (Again, just 100% minus the number in the previous line)
And we know that:
  • Amount we're risking when shoving = $1,500
  • Amount we win if everyone folds = $195 (dead money in the pot right now)
  • Average amount we could win if our shove gets called by one player = $1,590 (this is the calling amount plus the average dead money left in pot)
Next, per pokerstove:
  • Equity of our AK against calling range of one opponent = 18.5%
Finally, we apply the basic EV calculation:
  • 77.1% of the time everyone folds and we win $195
  • 22.9% of the time we get called by at least one player:
    • 18.5% of the time when we get called by one player our AK wins and we win $1590
    • 81.5% of the time when we get called by one player, our AK is beat and we lose $1500
Therefore:
  • Our jamming expected value = negative $61.74
  • If both opponents call, our EV is even less than this amount
  • If I loosen up both of the opponents' calling ranges further to, say, AK+ and QQ+, our EV becomes even more negative (-$103.69) 
  • Jamming all-in is therefore a mathematically bad play
  • QED
All-in for now...
-Bug