What Do You Do Different In Tourney & Cash Games?

[pzhon]

[ QUOTE ]
My apologies - my response was with SNGs in mind, and I should have qualified that. I don't have enough MTT experience to talk about those.

I was also assuming that the player in question was the best at the table (I don't think it's a stretch to think that a 2p2er is the best player at any given SNG, at least for lower limits).

[/ QUOTE ]
You might be surprised at how many 2+2ers are bad players, and how often 2+2ers run into each other (in which case they can't both be the best). Many 2+2ers sacrifice win rates in order to play more tables, which means they routinely get outplayed by people who are just playing 1-4 tables.

Anyway, SNGs are among the most distorted tournaments, where play differs from cash games the most. The portion of the prize pool awarded for surviving the bubble of a MTT is usually 10-20%. In a 9 or 10 player SNG, 60% of the prize money is awarded for surviving the bubble.

[ QUOTE ]

I'd be curious to see the calculations for QQ v AKo in the early levels of a SNG - I don't think I've folded QQ pre-flop in an SNG (I only would in weird bubble situations), but that's because people will donk off their stacks with all sorts of terrible hands. It seems obviously right to me to fold QQ if a guy pushed all in with a face up AK when there were only $30 worth of blinds in the pot with each person sitting on $1500 stacks

[/ QUOTE ]
Ok, let's go through the SNG math.

Let's suppose it is the first hand of a $30+$3 speed SNG on Party. You post the big blind of 40, and everyone folds to the SB who pushes his whole 2000 stack in.

If you fold, you have 1960 against one player with 2040 and 8 players with 2000. Ignoring your position (posting the small blind is actually pretty good, better than the hijack, but the blinds are small), the ICM says this is worth $29.45.

If you call and win, you have 4000 against 8 players with 2000. The ICM says this is worth $55.33.

Ignoring the tiny chance of a tie (I've checked, and ties have a negligible effect), this means you should call when you win at least 2945/5533 ~ 53.23%.

If you ignore the big blind, you should call when you win 3000/5533 ~ 54.22%. The big blind lowered the calling threshold by about 1%, close to your blind's portion of the pot if you call. One percent is far from negligible in this context. On PokerStars, the blinds on the first level are only about 2/3 as large, so they have about 2/3 of the effect, which is still significant.

Your skill advantage may raise your calling threshold slightly. For very strong players, it could be that folding QQ versus AK maximizes E$ in this tournament, since the SNG structure makes it a closer decision, and very strong players might expect to double up 55% of the time anyway. However, busting out lets you play another tournament, so I'm pretty sure a winning player will maximize his $/hour by calling. Calling just has to lose less than about half of the value of getting to play another tournament.

Again, in MTTs, the prize structure does not reward surviving nearly as strongly, and it would take very special conditions to fold QQ vs AK.

[jcl]

In The Mathematics of Poker there is a section on 'the theory of doubling up'. I was lent it by a friend so I don''t have the specifics but I remember that it concluded that you need a ridiculously large edge to pass up QQ v AK (in an MTT). Just another source you may want to refer to.