This might be a little off topic but I figured the smart guys in here would have the best answers. I created a thread addressing rake for OFC/Pineapple.

http://forumserver.twoplustwo.com/21...apple-1397727/

Suppose you start a hand with -- / A5 / KT6 and your opponent has -- / JT / K77.

On what street do you abandon your flush draw and pair up?

Click here to read the full analysis by Warren.

I thought this was an interesting situation.... it looks like the simple answer is to make trips on the bottom, but is it really that easy?

If we make trips on the bottom, we can try to go for front row royalties (4 or 6 points). If we do pair the top, we will scoop and get royalties. BUT if we don't pair the top, we will likely miss out on the scoop since it will be hard to beat his king high.

If we put the ace on the top, it looks like we will likely scoop but we are giving up on royalties. Which is better?

Playwarren?

EDIT: Missed that if we put the ace on the bottom it's also possible to hit the case king in the middle for a middle row royalty.

I'm playing it up top

Sephiroth,

It's fairly close, but placing Ace at the bottom is more profitable due to the front row royalties, lower rate of fouling, and a tiny chance of making royalties in the middle (trips+).

Thanks.

No kidding.

The Ace wins the front row leading to a likely scoop. It's impossible for villain to retake the lead in front row barring a one-out A which will hit him only 7% of the time. The middle is locked up and the back is in the lead, with redraws if villain hits two pair. And a two-out 10 is less than 15% likely to come.

After setting the A in back row, it's only ~32% likely to make the royalty in front, and only 38% to win the row outright (adding in the case A to the J9 outs). So, setting the A in back row is essentially a 5 point concession. It does absolutely nothing for the hand in the back row beyond firming up a lead in a row that is already won (and the K trips probability is also a low 7%). It adds no royalty value and devalues the scoop ev. However it does absolutely everything for the hand in regards to the scoop when set in front.

Villain improves back row about 42% of the time, but you also redraw about 32% of the time (with the A in front). So those two event probabilities are similar. Even if villain wins the back row, it's impossible for him to win more than 1 point unless you foul; the hand would have the same exact points outcome if you put the Ace in back and (basically) concede the front row -- a two-row 1pt win for you. A case could be made you're on a freeroll for front or middle row royalties, but it really costs 5 pts. Those points may be consequential or inconsequential. It depends on the situation.

So did you ever want to book a bet yet with me vs Warren?

We can't bet with you due to legal reasons but I'm sure I can find some people to xbook against you. You'll have to wait until we finish Supercomputer mode.

In the mean time, how many minutes can Warren think when figuring out where to place his card?

Are you open to playing Pineapple and Criss Cross as well?

Prefer to play criss cross or pineapple. Single hand ofcp not degen enough

How many minutes does Warren have per street? He obviously can't take an hour.

Why would a computer need more than seconds? I'm not going to take longer than 15 seconds per decision

In Supercomputer mode, Warren thinks by running simulations so he can determine the best play. Warren won't take that long in late-end games since he doesn't need to run that many simulations to have a high level of confidence. However, he'll need a lot more than 15 seconds in early game.

If Warren is weak, it doesn't matter how long he takes, his plays will suck and we'll call it a day within the first few games and not waste our time. I'm confidently this is unlikely to be the case so it'd be nice to have a least a few minutes per decision. He doesn't necessarily have to use all that time. You can use the same amount of time if you want.

Lastly, if Warren can only have 15 seconds to think, let's just drop the discussion right now and not waste each other's time.

Ya I'm not going to wait 1-5 minutes for a computer to make a decision on multiple streets I'd rather shoot myself. This is an action game

Some ppl are cheaters without admitting it
Fyi

even throwing out the number of 15min per action is pretty ridiculous imo.

Saw a play warren simulation on 12th street was zero % of foul but is said .5% to foul

Why?

Also wanna change my play warren password, but it doesn't show up anywhere

Emailed them, no response

You shouldn't be paying for that IMO
We've pointed out multiple flaws in it and they're still charging absurd prices

Oh damn didn't realize
They sold me well
Good marketing I guess

Thank you

A couple opening questions-

Q849T

I set this as Q/8/49T figuring that I can maybe back into fantasyland somehow and if not Q-high is an OK backup hand up top. Since I don't really want to pair the Q early anyway so this doesn't cost much, right?


K284J

Total garbage hands like this I usually put the 2-flush on the bottom, but it gets me in trouble when I don't hit a K or clubs. Is it better to just play it safe and stick with big cards?

Can you email us the screenshot of the hand in question? It should have the correct fouling %. Thanks.

The multiple flaws in question were on the initial FREE version released 3 months ago. We acknowledged the flaws and have fixed them. If there are more flaws that we don't know of, please email us or post them here for everyone to see. We want to fix these flaws to offer a better product for our loyal customers.

Also, it's pretty amusing that you keep slamming a product's prototype even though there have been massive improvements since the initial version and you likely haven't use the most recent update. IBM must be thanking God right now that you didn't get to play against Deep Blue's initial prototype.

For this answer I am assuming heads up.

1) Seems fine. Keep in mind as you set that you are probably drawing to 2 Queens, not 3, since a Q will almost always go into the flush draw. 3-flush OOP is ~55% and ~63% IP if totally live. If you don't like hovering around 60% you could also consider setting the 8-9-10 combo in back, giving you multiple ways to support a middle-row KK or AA and still catch some straight royalties. You lose the flush ev but it really depends if you like ~60%. Once you set for the flush there is no plan-changing.

ActionDJ or Hautari might have more to add on direct and indirect outs if you set up the 8910.

Q-high heads up certainly can be a winner, IMO that is dependent on how experienced your opponent is.

2) It's a raggy-looking hand at first glance, but the positive is that the K-J and 2-4 can be grouped in mostly separate drawing ranges so two-pairs and straights for both are possible without getting in each others' ways. You could have a decision to make with an ace.

2-flush, unless you're playing pineapple, hovers just around 30% to get there (less if OOP, more if IP and completely live). So, yes, K2 is easy to get into trouble with if you don't run the 70%-unlikely path or spike a K (which is under 50% regardless of position).

I might set 8/24/KJ to keep two-pair and straight draws live, maybe set any Q that comes in front (unless it completes a back row draw) and hope for QQ or 88. I think this is one of those hands you probably should be trying to lose the minimum.