Originally posted by MrMardyBum
Maybe this will help you decision....
Cue the Holecard profit table.....

Make sense now?
It's strange how I was about to start a thread about whether pushing AK might be -EV preflop on medium limits that this came out...
In response to the picture above, my observations through 85k hands on different limits are similar, with AK being aprox +1.7bb/100 hands. However, taking NL100 alone (20K sample), only AKo is positive at 1.2bb, whereas AKs is -0.2bb, at least proving that this hand has well above average variance.
I noticed that you take the hands as played both pre- and post- flop. For me at least, especially interesting part was how does it perform purely PF, that way filtering out the effect of potential postflop leaks. To my surprise (well, not really, seing how I was being hammered on every coinflip), on NL100, my AKs stands at -3.4bb, whereas AKo has -1bb. Sample sizes for PF pushes were 28 and 58 respectively, so I'm not sure what to make of it. Is the sample just too small (it still proves that AK variance is just GROSS), or might it actually be that SSS SHC for NL100 upwards should be adjusted against going broke with AK? With comparable sample size, it used to be hugely profitable on lower limits...
Would be grateful for your insights as well. Thanks!