sklansky bucks is not just about all ins. it computes EV of your bets on each street for all hands for which there was a showdown.
But yes, it does not include coolers.
I think Sklansky bucks are flawed though, mainly because the outcome of the board sometimes is what causes the hand to go to showdown or not.
Basically let's say villain has a draw, you bet turn, he calls, whether his call is correct or not is irrelevant. Now if the river blanks he loses but if the river hits he commits more money to the hand. Now according to Sklansky bucks, in this situation your villain will hit 100% of the time, thus making you unlucky (even if they were a favorite in the situation when they called, they got more than their
equity). But when they miss, they fold and it doesn't count you as lucky. Obviously it happens the other way as well, but since in general most people here will have opponents calling bets without implied odds more than they will be doing so themselves, I predict that all good players should have more Sklansky bucks than real dollars in the long run, simply because of this one flaw. All-in ev however is independent of this because once a hand is all-in, you see both hands every time and the same amount of cards will be seen every time.
Example:
Hero has AKd, villain has 35h.
Hero raise to 4 big blinds, villain calls - 8 big blind pot, 5 are heros, 3 villain's.
Flop comes AhKc3s rainbow
Hero bets 6 big blinds, villain calls - 12 went into the pot here, 10 belong to hero, 2 to villain.
Turn comes Jh
Pot = 20, hero bets 15, villain calls - 30 goes into the pot, 23 belong to hero, 17 to villain
Now we deviate. Three possibilities:
River comes blank, hero bets any amount, villain folds.
River comes heart, hero checks, villain bets 25 big blinds, hero calls. All 50 here belong to villain.
River comes 3, however it happens, hero loses all his money.
44 cards left, 2 outs give #3, 9 bring #2, other 33 bring #1.
So technical
real money for last scenario would be -(2/44)*50 - (9/33)*25 = hero loses 9 big blinds.
Overall Sklansky bucks for hero assuming we see hands every time: 5+10+23 - (4+6+15) - 9 = 4. This is profitable for hero.
But since it doesn't always go to showdown, Sklansky bucks when it goes to showdown: 5+10+23 - (4+6+15) -(2/11)*50 - (9/11)*25 = -16.5
Money lost when it goes to showdown: (2/11)*100 + (9/11)*50 = 59.
So over the long run, making what is a good play with positive sklansky bucks were all the cards showing, yields a result such that over the long run (meaning luck is 0) your Sklansky bucks will be -42.5 big blinds (yes almost half a buy-in) per time this happens. That's a pretty big effect, and all because it neglects to realize that oftentimes whether a flush or straight hits depends on whether the draw hits. So this means that using the Sklansky bucks method, practically every time a hand of this type reaches showdown, the draw will have hit, and it will look like those drawing (even unprofitably) are hitting their draw every single time.