Computing odds in Video Poker

I'm on the plane, heading to Vegas for G2E, trying to sleep so I don't suffer jet lag, and I can't keep my mind off casino games. Go figure...

So I gotta pull out my iPhone and start typing even though i cant post until I land. Am I compulsive about WoV or what?

Anyway, i've got video poker rattling around my head, wondering if stupidity is taken into account when the odds and/or house edge are calculated.

What I mean is, if a rational player is dealt two pair, he's gonna discard that one extra card, knowing the best he'll do is draw a full house.

But if an idiot was playing, and got dealt for example, 6 J J K 6, he might see only the jacks. Heck, on a JoB machine, he might even be happy and intentionally hold only the jacks.

Now the best improvement would be quads. I admit it's rare, but does the math take it into account?

What about when you're dealt a hand such as 9 10 J Q K where the 10 J Q K were suited? As has been mentioned in other threads, you should toss the 9 leaving 1 chance for a Royal, 1 for a SF, 7 for a flush and 5 for a straight. 

Is THAT type of play considered?

I'm thinking the math guys need to consider more than the simple scenario of 5 cards plus draw. I.E. If the ten cards, in order are: A Q Q Q 5 A A A 7 4, a Basic Strategy player would toss the A and 5, only to get two other aces for a boat - but immediately wonder what the next 2 card were... But if you consider the "idiot factor," then the best possible hand Is quads. On the other hand, if the cards are: A Q Q Q 5 7 4 A A A, then the best hand is trip aces, because if the idiot holds one ace, he'd never see the last ace, since he'd only draw the 6th thru 9th cards. And if the idiot discards everything, then he'd still draw trips. Am I making sense?

Bottom line: Do the math guys take idiots into account? If so, how accurate are posted returns for a Basic Strategy player?


FWI: I rarely play video poker, so I have no idea why my mind decided to wander in that direction.

I gotta get some sleep....

You have questions even though your rarely play video poker? So here's a comment from someone who NEVER plays the game....

I think the calculated returns are based on making the "optimum" play each time. This requires that ALL of the possible holds be compared first, to determine which is optimal, even considering the play to discard BOTH pair and hold an odd singleton. Stupid, maybe, but you have to establish whether it is stupid before you can dismiss it and calculate the return on something that appears "smart" at first glance.

Quote: DJTeddyBear

What I mean is, if a rational player is dealt two pair, he's gonna discard that one extra card, knowing the best he'll do is draw a full house.

But if an idiot was playing, and got dealt for example, 6 J J K 6, he might see only the jacks. Heck, on a JoB machine, he might even be happy and intentionally hold only the jacks.

Now the best improvement would be quads. I admit it's rare, but does the math take it into account?

Uh, with two pairs, holding just one of the pairs is often the correct strategy. You would do it in Deuces Wild whenever Full House pays less than 4x and you would also do it any JoB variation where two pair pays 1x (not 2x) and quads have some extra large payoff.

So some of the plays you call 'stupid' may be optimal.

I don't know what the strategy error cost for an average video poker player is. When I see people play, I see them making mistakes but not horribly costly ones (the error I've seen most often is holding four to an inside straight). I could suppose that a VERY bad (or stupid if you call it) video poker player could make errors adding up to 10% house edge, and even he might get lucky when being stupid.

All the strategy guides I've read, including the Wizard's, specify that the return for a given pay table assumes perfect use of the strategy.

Still, I suppose it's possible the percentage for comps or cash back set by casinos may have something to do with how people play such games.

Drinking a Denver Pale Ale and waiting for the plane...

Yes, the odds and return for video poker are based on making the perfect play every time, which is the hold combnination that pays the highest average pay. It does not take into account the volatility of a specific play, although I have read that some will use volatility as a secondary decision if two holds pay identical.

See you in a few days!

Many jurisdictions factor in 3% for bad play. Some (I can't recall which, though), have an 80% min for slots and an 83% min for poker, although I've never seen a poker that bad.

basically on any hand youre dealt theres 32 different hold/discard decisions you can make. the payback percentage is the sum of the return picking the best of these 32 decisions for each hand you can possibly be dealt.

1 way to hold all cards
5 ways to hold 4 cards
10 ways to hold 3 cards
10 ways to hold 2 cards
5 ways to hold 1 card
1 way to hold zero cards

Doc -

Yes. Even though I rarely play it, I'm curious about the math. For the record, when I do play video poker, it's always JoB. Usually on a 50 or 100 play machine.


Jufo -

I did specify JoB. I was unaware that there's any version that pays less than 2 for two pair.

Quote: DJTeddyBear

Doc -

Yes. Even though I rarely play it, I'm curious about the math. For the record, when I do play video poker, it's always JoB. Usually on a 50 or 100 play machine.

I'd suggest you visit the website: www.vpgenius.com. It has all the possible video poker variations and optimal strategies for each. You can create you own custom paytable and analyze the strategy for it. So you could for example make flush pay 7x instead of 6x and check how the optimal card discard decisions change with that. The hand analyzer there will also show you the EV of each possible decision so you get to see how much each incorrect play costs you in expected value. If you play with numbers at that site a bit, your knowledge of the inner workings of Video Poker will greatly increase, I can promise you that!

Quote: DJTeddyBear

Jufo -

I did specify JoB. I was unaware that there's any version that pays less than 2 for two pair.

Ah, sorry about that. Even JoB has different paytables so just mentioning JoB is not ambiguous. But I agree that all variations of JoB probably pay 2x for two pairs.

As an additional note: In some cases you even discard pat Full-House and only keep the three-of-a-kind from it if quads pay high enough. This situation arises in some Double double bonus variations.

i always get a chuckle seeing 8/5 JoB and 8/5 BP on the same machine for the same denomination.