Sklansky hand rankings

[RevStu]

I was buying a card cover for a friend on eBay, which purported to have best-starting-hand rankings on it, and on close inspection noticed that the rankings seemed to be a bit off. A swift bit of research later showed that the cover was listing the Sklansky rankings, which go like this:

1. AA
2. KK
3. QQ
4. JJ
5. AKsuited
6. AQs
7. TT
8. AK
9. AJs
10. KQs
11. 99
12. ATs
13. AQ
14. KJs
15. 88

The site where I found the list said this:

However, these charts were created by Sklansky without any definitive proof of why certain hands were better - they simply were.

So what's that all about? The supposed 15th-best hand is actually a 53-47 favourite to beat the supposed 5th-best (AK suited), and in fact against the 15th-best hand you've got a better chance with the 14th-ranked hand than you do with the 5th-ranked one. Even the 71st-best hand (22) is favourite heads-up against the supposed 5th-best.

And since everyone here usually comes down like a ton of bricks when I suggest that you can't rely on probability to manifest over a time-span as limited as a poker player's career, how come one of the most respected writers on the game can get away with making this claim without a load of forum posters ridiculing it? What's the reasoning?

(I thought at first that perhaps the rankings were based on multi-player hands rather than heads-up, but if you start adding other hands then the odds for the 88 against the AK or AQ generally get better rather than worse. Against AQ, for example, adding random hands still has the 88 as favourite until you get to a six-way pot, and how often do you ever see one of those?)

[RevStu]

(The only other obvious thing is that even if you're sitting with 88 and it's still the best hand, you're probably going to fold it to betting on an overcard flop regardless. But if you fold every time someone bets with an overcard on the board you probably shouldn't be playing poker anyway.)

[RevStu]

Depends what the flop is...

I see what you're saying, of course, but if played properly 8s are still going to be a very strong hand against a LOT of flops, probably more than 50%. And IMO there's absolutely nothing more fun in the whole of poker than rinsing someone who just won't let go of a missed AK...

[Dice Man]

What Scrote said...

Even though 22 is a favourite against AKs, it's still a terrible hand whereas AKs is a premium. Why? Because AKs plays better against other hands. Remember if you have 22 vs 88 you are in worse shape than a player with A7 is against a player with AK. 22 is an underdog to most hands I think, whereas AKs is a favourite against most hands.

This is my hand ranking chart. (Not that I or anybody ever uses it)



So it seems Sklansky and I disagree about a few things, like, what's better: AKs or JJ, AQs or TT,. I'm also surprised that he thinks ATs is better than AQo. But it seems we don't disagree too much. I must over-value high cards to pairs and undervalue suited cards, in his eyes.

Also on my chart I value all the hands on the same row equally, or at least I haven't discerned any significant difference between them.

[PokerWoody]

Dice,

Please stop posing that. It makes little sense.

Rev,

Is he talking about limit or nolimit? For limit that chart looks pretty good, for NL no way. Also, comparing one hand to another with pokerstove and saying something like the 15th best hand is better than the 5th best hand is just thinking about it in the wrong way. Top 10% in each (from pokerstove) against each hand for example...

56.419% { 88+, A9s+, KTs+, QTs+, AJo+, KQo }
43.581% { 88 }

42.216% { 88+, A9s+, KTs+, QTs+, AJo+, KQo }
57.784% 52.64% { AKo }

Also, this is why I think pokerstove's top x% chart is wrong. It works for limit, but not NL much like these rankings.

[piquetb]

I was buying a card cover for a friend on eBay, which purported to have best-starting-hand rankings on it, and on close inspection noticed that the rankings seemed to be a bit off. A swift bit of research later showed that the cover was listing the Sklansky rankings, which go like this:

1. AA
2. KK
3. QQ
4. JJ
5. AKsuited
6. AQs
7. TT
8. AK
9. AJs
10. KQs
11. 99
12. ATs
13. AQ
14. KJs
15. 88

The site where I found the list said this:



So what's that all about? The supposed 15th-best hand is actually a 53-47 favourite to beat the supposed 5th-best (AK suited), and in fact against the 15th-best hand you've got a better chance with the 14th-ranked hand than you do with the 5th-ranked one. Even the 71st-best hand (22) is favourite heads-up against the supposed 5th-best.

And since everyone here usually comes down like a ton of bricks when I suggest that you can't rely on probability to manifest over a time-span as limited as a poker player's career, how come one of the most respected writers on the game can get away with making this claim without a load of forum posters ridiculing it? What's the reasoning?

(I thought at first that perhaps the rankings were based on multi-player hands rather than heads-up, but if you start adding other hands then the odds for the 88 against the AK or AQ generally get better rather than worse. Against AQ, for example, adding random hands still has the 88 as favourite until you get to a six-way pot, and how often do you ever see one of those?)

Ranking against what ? Its pointless comparing hands unless you can see opp's cards. You need to compare your hand against his range and this will depend on a number of factors

eg number of players

Against a random hand 88 is 69% and AKs 67% but this reverses with even one more player. First in at a full table, AKs is 22% and 88 only 14%. ie >50% better, even without playing better on the flop. So heads up and ring situations are totally different (like the demerits of 32 and 72)

eg how tight is opp

Against a random hand AQo and ATs are about the same. Against a super tight opponent ATs is 30% against AQo at 28% but this is the only time its ahead (heads up anyway, Against an average opponent AQo is 50% and ATs only 40%. As HU is not super tight AQo is better here.

eg position

If UTG in a full ring game comes in with a standard raise you're playing AKs, but you might be folding 8's (unless you can afford to play farmer)

[RevStu]

Against a random hand 88 is 69% and AKs 67% but this reverses with even one more player. First in at a full table, AKs is 22% and 88 only 14%. ie >50% better, even without playing better on the flop. So heads up and ring situations are totally different (like the demerits of 32 and 72)

How often are you going to have 8 or 10 players actually in a hand, though? Basically never. Most pots get played out between 2-3 players.

I get the point that odds change with different numbers of players, and beyond that with different types of players, but since there's no possible way of judging those qualities, which are unique to every situation in every game, surely the only even slightly meaningful way of coming up with a ranking is to rank the hands directly against each other?

(My personal take, for example, is that in the case of 88 or any other underpair vs AK, the "real" odds for AK get worse the more players are in the hand, because people are more likely to be playing As and Ks, so you're going to have fewer outs against the pair than you would in a heads-up hand.)

[piquetb]

How often are you going to have 8 or 10 players actually in a hand, though? Basically never. Most pots get played out between 2-3 players.

I get the point that odds change with different numbers of players, and beyond that with different types of players, but since there's no possible way of judging those qualities, which are unique to every situation in every game, surely the only even slightly meaningful way of coming up with a ranking is to rank the hands directly against each other?

(My personal take, for example, is that in the case of 88 or any other underpair vs AK, the "real" odds for AK get worse the more players are in the hand, because people are more likely to be playing As and Ks, so you're going to have fewer outs against the pair than you would in a heads-up hand.)

Take AKs - three players call with top 10% hands (actually 9.8% on Pokerstove) - you win 29% (post flop considerations aside). With 88 its only 22%. So AKs is better in real life in a ring game.

Some hands beat hands that beat hands that beat the first hand (Roshambo style) so how are you ever going to get a definitive list (and why ?)

[lev1athan]

Ranking against what ? Its pointless comparing hands unless you can see opp's cards. You need to compare your hand against his range and this will depend on a number of factors

eg number of players

Against a random hand 88 is 69% and AKs 67% but this reverses with even one more player. First in at a full table, AKs is 22% and 88 only 14%. ie >50% better, even without playing better on the flop. So heads up and ring situations are totally different (like the demerits of 32 and 72)

eg how tight is opp

Against a random hand AQo and ATs are about the same. Against a super tight opponent ATs is 30% against AQo at 28% but this is the only time its ahead (heads up anyway, Against an average opponent AQo is 50% and ATs only 40%. As HU is not super tight AQo is better here.

eg position

If UTG in a full ring game comes in with a standard raise you're playing AKs, but you might be folding 8's (unless you can afford to play farmer)

It's not practical to include factors like position and opponents when using stats like these - it would require so many extra dimensions of complexity as to make it unusable. For instance if it folds round to you 9-handed with 72o on the button you're in a better spot than T7o UTG, but no-one's saying 72o is a better hand than T7o.

These things should only be regarded as tools, not idiot-proof guides; and the hands are only comparable when the other variables are the same, as with any test.

I have to agree with Dice that ATs being "better" than AQo jumped out at me as being weird too; would be interesting to know how he came up with this.

[swinebag22]

I have to agree with Dice that ATs being "better" than AQo jumped out at me as being weird too; would be interesting to know how he came up with this.

It could be the soooooooooted factor, but could also be the straight factor, of which most have a T, but agree that it is a strange one

have only skimmed the thread, but are some people getting confused between starting hands and preflop showdown hands?