[pokergripes]

I think the point of a raise to 15k is four-fold (the first of these is the original poster's idea, not intending to take credit for it):

--First, even if I'm not folding on a flop of, e.g., AKQ rainbow with no heart (because I could be beating 2-3, might hit an eight or six on the turn or river, have far too much of my stack committed, etc.), the typical player does not want to take that chance by just calling my 15k pre-flop, and will therefore feel some pressure (even if it's analytically incorrect, since it's not his job to bust me) to make sure I'm busted if I lose...nobody likes the "chip and a chair" thing to make them look dumb in front of their peers.

--Second, 15k indicates more strength than all-in (looks like you're inviting a call), and you might pick up a bit of fold equity that way.

--Third, it looks odd enough to some players that it confuses them, and therefore might make them less likely to think it through clearly (and clear thinking would lead to a call for sure from the bb with pretty much any two if you push, assuming the sb folds).

--Fourth, if the first point above doesn't result in me getting set all in pre-flop (that is, neither the sb or bb who calls is the kind of opp who thinks it's his job to bust me), then it also could get checked down after the flop if the caller totally misses (obviously pretty unlikely, but not 0%), and in that case I retain ten antes and a chair.

I'm not saying the 15k is the "right play", but rather, that all the choices are pretty grim in this situation, the push-versus-fold decision is on the cusp (strong views in this thread that it's "obviously" a push or "obviously" a fold notwithstanding), and when something is that close to a toss-up, doing something different might be better (and a flat call with the six eight is clearly bad news).

On the other hand, I was pretty drunk when I asked the question at 3am this morning, and now lean more towards the push than the 15k (after reading the responses), above points notwithstanding. But I think they're all pretty close alternatives (assuming randomized players and no prior hand experience with them to help the analysis).