Pokerstars opened up 10/20 badugi a couple days ago and I'm addicted. I was playing around with starting hand strategy today, but I'm still sorting out relative strengths of 2-card, 3-card, and badugi starting hands. When no one is pat, it is so important to have the best three-card hand (unlike in triple draw, if you and your opponent draw 1 and both miss, whoever had the best draw always wins; so e.g. that final draw is like 80/20 instead of 55/45 or 60/40).

Getting there is by far the most important badugi skill, but 2a and 2b seem to be not missing bets when you have the best 3-card and avoiding getting stuck in pots with second-best (or worse) 3-card, especially in headsup pots. If people have good snowing frequencies, the pat vs draw situations get more interesting, but the Stars game doesn't seem to have much of that.

Only ~6.5% of hands are dealt badugis, and 3-card 6s or better are only another 6%, so while a full-ring EP strategy (I think Badugi is usually dealt 7-handed, though the Stars game is 8-handed) of folding everything but badugis and strong 3-cards makes sense, in late position and in shorthanded games it's pretty obvious that's too tight.

The question is how to compare good 2-card starting hands and rough 3-card starting hands. Clearly A23x > A27x > A2xx, but it isn't clear to me where, say, 654x should fall. It's a lot harder to turn a 3-card into a badugi than it is to turn a 2-card into a 3-card. Even though 765x will make more badugis than A2xx, A2xx will end up with a lot of better 3-cards, and 765x can't really promote itself to a better 3-card because it's so rough. On a single draw, A2xx has 10* outs twice to improve to beat 765x, and 765x only has 10* outs once to improve to a badugi, and a couple more to notch A27 (but none of the other three-cards A2xx can make); with three draws it has to be the equity favorite. It's not that hard to figure out a one-card draw's equity versus a pat badugi. Finding, for example, the weakest 3-card hand that is favored over A2xx would be a lot trickier, but the general problem of comparing 2-card hands to 3-card hands is something I want to work on.

* - I didn't account for card removal (blockers in the other hand) there, which affects the 2-card hands improvement chances a bit more than the 3-card's.

A bit of trivia: A2xx is the single most common starting hand. Any specific badugi is ~ 0.1% of the dealt hands, A23x is the most common 3-card at ~0.33%, and A2xx is ~1.3%. (The best two-card and three-card hands are more common than worse hands of the same type because you always discard the worst suited card e.g. As2c3d4d is played as A23 not A24.)

912k vpp and counting...

Flop texture is too big a topic to do it justice in a couple paragraphs, but if we ignore some nuance we can separate PLO flops into three big categories: dry flops, draw-heavy flops, and obvious-nuts flops. The archetypal dry flop is K82r; there's no way besides sets and pairs to hit it, and there are relatively few hands that hit it. There is a wide range of draw-heavy flops but a good example is Th9h5s; there are a ton of ways to hit this flop, including a variety of wraps, combo draws, and pair+draw hands, in addition to sets and top two. Obvious-nuts flops include monotone flops and super-connected flops like 654r. They look a lot like the draw-heavy flops, and in holdem they play essentially the same because in holdem the nuts isn't very likely and all of the obvious-nuts flops also contain a lot of pair+draw possibilities. But in PLO, the nuts is much more likely (and on monotone flops, there are no one-card flush draws), and as a result those flops share some characteristics with the K82-type flops. Specifically, both of those flop-types differ from the Th9h5s flop-type in that there are many fewer hands an opponent who doesn't have the nuts can easily be aggressive with. The strongest semibluffing hand on a monotone flop is a set, and it's a solid dog to a flush; the strongest combo draw on a 6s5s4c flop (flush draw+gutter) is an 11-outer versus 87.

What we're really talking about here is the distribution of equities among the hands that hit the flop reasonably hard. What separates Th9h5s is both that a wider range hits it, and that if you run equity calculations matching up individual hands in that range with the entire range, you get a narrower distribution of outcomes. In other words, there are a ton of hand vs range 55/45s on that kind of flop, and although there are individual hand versus hand domination situations (set over set, dominating draw) there really aren't any hands strong enough to dominate an entire range. At the other extreme is K82r, which doesn't have any semi-bluffing hands and where KKK has massive equity versus any range. Stretch from the driest flops to the slightly drawier dry flops, like AJ4r, and similarites emerge with flops like 654r and monotone flops; there are fewer combinations of hands with decent equity against the nuts, and there is more separation among relative hand strengths.

These categories are too restrictive and the dscussion above isn't sufficient to cover all important points. For example, a main difference between AJ4r and 654r is that there are 3 available two-card combinations of AA on the former and 16 of 87 on the latter). Also, flops like Th9h6s have a very vulnerable flopped nuts and are essentially a slightly different kind of draw-heavy flop. That said, just understanding this basic distinction will help with some flop decisions. Some of the same factors that may make K82r a good flop to cbet or to bluff raise apply to 654r. And you may find opponents are predictable in similar ways on AsJc4s and 6s5s4c; moderately passive opponents don't semibluff these flops much, and check-call a lot of decent hands that they will fold later.