7 cards, and just half the bet. Intuitively, I knew it had to be worse than the "normal" game, but I don't know how often the dealer actually gets a K-high nothing or worse under normal circumstances (when I'm at the table, the frequency of such hands converges to 0 pretty quickly).

Obv the situation where this is actually advantageous to the player (dealer has K-high, player has Q-high or whatever) is damn near infinitesimal. OK, pretty rare...

The dealer was explaining it to me, and when he got to the half-bet part, I started laughing and said, "That's a pretty big commission!" He said, no, you get half whether you win or lose!

I'm sure there's been a time where I had a J-high vs a dealer K-high, but I don't remember it.

It's an automatic half unit win on our total bet when the dealer has K high or lower in the back? Always? Even when we would have lost the front? Because in these scenarios we lose the front a little more than half of the time anyway, so gaining a full half unit in this spot is a good thing. If the rules are this way, this could be better than normal Pai Gow. But I think I'm misunderstanding the rules.

The dealer gets a K-high or worse 6.71% of the time. According to my calculations based on the Wizard's numbers, a player would beat the 5-card hand of a dealer at the Trump Plaza who had a K-high or worse 99.75% of the time. This doesn't account for the 2-card hand at all, but the highest a dealer could have is QJ, and the player has a 67.4% chance of having Kx or higher in the 2-card hand (without any removal effect, given the dealer's hand range constraints).

So we win the back 99.75% of the time, and win the front AT LEAST 67.4% of the time (since I'm not accounting for beating the dealer's 2-card when we have QJ or less). Having to accept only a 0.5 bet win here isn't good.

However, when we beat the dealer when he has bigger than K-high, we win 1 full bet, not 0.95 bet. This only happens 28.6% of the time though (actually less, since that figure includes hands when the dealer has K-high or worse).

This is also when the player never banks (which might not be possible?). Just looking at these rough numbers, it doesn't look like a good variant for the player.

Wow, I'm a tard. When he has a king high back, the highest front he can have is QJ. But I'm not sure where you got the player beating a K high from the dealer 99.75% of the time. If the dealer gets a K high in the back 6.71% of the time, so do we. And when we do get K-high or less, we win half of them, barring a nearly impossible front copy with a high card. So our overall equity is closer to 100 - 6.71%/2 = 96.645% in the front.

Now for the back, like you pointed out, the dealer must play QJ or less. Here the best way to determine the dealer's average front is to start counting out all the K-high hands or lower. But since I want to watch overtime hockey, I have to come back to this. From rough inspection though, I don't think the house edge changes a lot.

http://wizardofodds.com/games/pai-gow-poker/appendix/1/

For 5-card hands 9-5 through K-10, I multiplied their probabilities by the Banker PR stat, then summed the products. That came to 0.25%, which represents the probability of the dealer getting K-high or lower AND beating you in the 5-card hand. The top of that page explains what those figures mean.

Yes we get it 6.71% of the time too, but given what we're comparing, it's only relevant if the dealer also has it (6.71% of time). The chances of both of us having K-high or worse is 6.71% * 6.71%, or 0.45%. Cut THAT in half, and you get 0.23%, which matches my figure above when you account for rounding and banker winning ties.

edit: these are not hyper-accurate numbers, since I'm not accounting for the effect of card removal... but they're ballpark figures.

Thanks for the info, guys.

I'm not big on conspiracy theories, but any of you ever dick around on the play games at Bovada? Would they really do something like this on purpose in order to encourage real-money play?

I'm thinking serious programming error -- I recall looking at that game, and what I saw was accurate for their rules.

Wow how does the banker not play J4-65432 there?

banker plays the flush with 43 on top.

Most casino house ways would play J-4. If they can play both a straight and flush, they play the one that yields the higher 2-card hand.

Yeah, house way has always been to maximize they low when they have a choice between pat hands anywhere I've played. Would be interesting to see if one could reverse-engineer where the error is and determine how it affects the house edge.

My guess is that mistakes like this happen because the house writes rules like this:

As a programmer, I'm pretty sure I'd have to fumble with the code a few times -- and assume I properly parsed the meaning of the instruction in the first place.

That appears to be the rule the programmer began to follow.

Borgata now has a no commission Pai Gow table. The catch is that if the dealer has queen high, every hand is a push. Does anyone have any stats to show if this is better or worse in the long run than paying a 5% commission.

That's very interesting. I can't imagine they would offer it if it didn't have a higher house edge. Or possibly it's neutral but the game moves faster because they don't need to use rolls of quarter, so they make more anyway and game is faster. Would be interested in the math.

Its called EZ Pai Gow and has about a 2.5% HE. That is around .25% better than being a player a regular Pai Gow BUT you can't bank and so the HE is actually much worse than standard Pai Gow if you bank with any frequency, particularly at a table with multiple bettors.

The inventor regularly posts over at the wizardofvegas forums(spinoff of wizardofodds). He is a dealer and strongly feels banking at Pai Gow is archaic and also a reprehensible behavior and so invented this game to eliminate it.

Obviously I know of systems where you add up the value of your two pair and keep them together only if you have a certain high card to put in your low hand. Anyone know of point systems for situations where, say, you have two pair plus a joker? Or a hand where you have, say, a flush, a straight and a pair?

Had a hand where I had 8-8-7-7-Q-2-Joker. Tried setting it as 8-8 up top and 7-7-Joker-Q-2 on the bottom. Other players said house way is Q-Joker up top and 2 pair behind. Would've won either way. Just situations like that one sometimes throw me off.

Like where is the breaking point where the above example becomes pair up front, two pair behind instead of pair up front, trips behind?

You must play in California, where the Joker is truly wild. In most casinos, in that scenario the Joker would only be an Ace, so you actually fouled your hand by setting it that way.

To answer your question though, I have no idea. Since I play on the east coast, I never studied up on "truly wild" Joker strategy.

Correct. San Diego to be exact. On one hand, it sucks b/c all the apps for pai gow poker I've found for Android are where the Joker is only used for an ace, straight or flush. Obviously the situation I listed and others like it are quite rare and the advantage one way or the other is very slight at best. Just found that hand example and others to be a bit confusing and sometimes counterintuitive to me. Another example I can think of is a hand like A-Q-T-9-7-5-2. Think house way in most casinos is to play Q-9 up top and A-T-7-5-2 down low instead of Q-T up top and A-9-7-5-2 down low. Guess because Q-T and Q-9 are essentially the same and if Q-T were to be good up top, then Q-9 would most likely to be and to strengthen your ace high down low.

Nice bump - was playing last night, fanned open a hand (instead of squeezing) and saw a bunch of middling clubs + joker. Heart skipped a beat, then I saw the 5 :-( I had J9875 + joker + 5. Guy next to me had the 6, of course.

Assume you played Joker-5 up top and Jc-9c-8c-7c-5c down low (also assuming that this is at a place where Joker can be anything)?

Where I play, they have no bonus or dragon bet but have jackpots for certain hands. If player has A-A up top and at least four of a kind, they get at least $100. Depends on the four of a kind and if they have a Joker or not. Think it's like $500 for A-A up top and four kings or better w no Joker. Also have jackpots depending on banker's hand. Most of the lower ones are 100% to banker. Yet if banker has A-A up top and four kings or better w no Joker, than banker gets 50% of $5000 and players who paid comission split the remaining 50%.

Was playing and was only one of two playing. Dealer had A-A-8-8-8-Joker-x. For that, players got nothing but if they had the last 8 instead, would've split $1000 I think (think it was $2000 for A-A up top, any four of a kind w no Joker, with players spliting 50%)

Not a fan of fanning my cards. I'm a squeezer. Adds lot of excitement and a sweat to the game.

Joker = A/Straight/Flush, so 5 5 / SF. I figured it was at least good for a push.

I'll squeeze if there's slow people at the table and I get my hand early. If there are really slow people, I'll look at the cards one at a time when they're face down or some other stupid crap while they try to figure out WTF to do before asking the dealer. Always fan if I'm playing 2 spots, though.

Of course. Didn't see that you were open ended to the J-high straight flush. Only saw the gutshot to the 9-high straight flush.

Where I play, you can only look/set one hand if playing two or more spots. Assume it's the same where you play?

Where I play you can set 2 hands, but have to play house way on a 3rd hand if you play that. So it varies.