Version 0.5 available here

Now with triple draw support. Please note I have BARELY played TD, I hope the method I provided for specifying drawing behaviour is usable. Any feedback is welcome, if you think there are better ways to specify how someone will decide what to keep or discard between draws, please let me know.

As with badugi, I know it would be better if it could account for betting action / opponents' drawing actions etc, but that's a whole new and much bigger problem, which I may eventually get to.



For hands that are drawing, you can specify
- what to stand pat with
- what non-wheel cards you are willing to keep if they improve your draw (in general only applies when drawing 2 or more, the exception is a 1 card draw that might make a straight or flush) - this is how you declare you would be willing to keep a 6 or an 8 or higher card
- which straights you would prefer to break by discarding the highest card. By default it will auto discard the 2nd highest card. This only needs to be specified for 8+ high straights, 6 and 7 high straights auto discard the 6.

Flushes are resolved by discarding the highest card unless it is a straight flush (in which case it is handled like a straight) or a wheel 75432 (it will discard the 5).

Also I fixed the badugi bug mentioned previously, it might affect some sims involving 2 card hands by about 0.2% or something

Finally there is also now a forum-friendly text output option with some suit symbols and bold text on final EV % values.

awesome. thanks

a few suggestions after 5 minutes of playing with it.

a) for 2nd and 3rd draw, have a checkbox or something where you can toggle whether to draw 2 or not. because most people will fold rather than draw 2 on 3rd draw basically no matter what.

b) this is more of a UI thing, it would be nice to get some sort of breakdown on when your hand got there. like, right now you have win/lose/tie, another nice stat would be how often and when you made your hand, like
made hand after 1st draw = x%
made hand after 2nd draw = y%
made hand after 3rd draw = z%
not sure whether it should be cumulitive or not.

also a stat like won% when made hand, so you know whether you lost because you bricked or lost because you lost to better made hand, or won because opponent bricked, etc.

also might be nice to have a range of what hands you ended up making, like

75432 - x%
76432 - y%
....

--------------
super cool though i'm sure the td experts who already have this kinda software are really bummed this is out there now lol.

also would be nice if you could choose between the current td setup which is all 3 streets, and just a one street td simulator.

Well you could set the 'keep' box to 689tjqka, i.e. include everything up to an ace - actually I didn't document this but that bit uses the same number parsing that is used for stud pair/kicker ranges, so I think you could put "6,8+" in that box and it will draw to as bad as ace high. Would you ever want to go worse than that, e.g. you have 77742, are you ever discarding only one 7 and drawing to a pair?

Oddly enough I originally had it dump out a bit more information, for showing win/lose tie % for a final hand of 8/better and 9/worse, but it seemed a bit arbitrary. I guess I should do something like this though, ideally I'd like to do it without cluttering up the UI with too much text. I did this in Badugi (result % for badugi, 3 cards, etc) so you can see how often you make your hand and still lose, and I suppose this is still important information.

See "number of draws" box, you can do 1 or 2 if you like. My tests with it set to 1 seem to match up with twodimes, which is encouraging as it means it's slightly less likely I screwed something up

cool i didn't see # of draws, you can set to 1.

but the main thing i was pointing out is that for simulation purposes, if a hand is still drawing 2 for the 3rd draw, to most accurately simulate game conditions, this hand should never be able to win. so by having a checkbox or something that says "don't draw more than" x, you can set x to say 2 for 2nd draw and 1 for 3rd draw and then you can start out with a one card hand, say 2c, and then draw 4 on first draw. you know, to simulate big blind or something.

for example, say you start with 237. now you're gonna draw 2 on first draw. even on 2nd draw you're gonna draw 2 most times, so the current simulator is ok here too. but on the 3rd draw if you still are drawing 2 then you will fold and never win the pot. but in your simulator 237 drawing 2 on last draw will win maybe up to 20%.

programatically, if you have checkbox set to never draw 2 on 3rd draw, you could just set the hand to 4 of a kind, which is equivalent to a fold since it will never win.

ok, i guess the problem is when both players are still drawing 2 after 2nd draw. maybe you could make that a special case, or maybe you wouldn't let all players set the option not to draw 2 on 3rd draw (which would simulate the button as the button generally would win all those hands).

Ok I know what you're getting at, and this is what I'm talking about when I say: The thing is, for someone to fold, I need to know that another player bet, so how do we know they are betting? Are they only betting when checked to, in which case, who has position? There is a huge difference between having the players play a virtual hand, complete with betting action, and just dealing the cards out to see how often hand A beats hand B.

Right now, this is supposed to be a hot/cold simulation. Here's a quote from the pokerstove FAQ and this is exactly what applies here: So while it would be more useful to figure out how profitable it would be to play a certain hand, currently the program is just telling you what the win % is if you are all in for your last bet predraw.

Anyway, I may eventually end up doing this but like I say it is a much bigger and more complicated problem than just figuring out win percentages, because in order to know that player A should fold, we also need to define how we know player B will bet.

This is amazing work, RubbishCards, congrats ! You're making a great addition for the community.

Thanks a lot for some very nice work, this have saved me some time and definitely helps getting insight into games!

I completely understand you not wanting to make a full-blown betting simulator, but there's probably some low hanging fruit that could make the calculations more like a betting simulator, as long as the user is aware of the limitations and use them appropriately.

A suggestion, if you are looking for ideas:

Graphing the equity from street-to-street like propokertools does. This helps a lot in distinguishing between implied odds hands and reverse implied odds hands.

I *think* adding a "fold if X on N-street"-function might go a long way of reducing boring work for the draw games, while being fairly simple to implement.

I don't really know how hard either suggestion is to implement it, but to me they seem conceptually simple and not as finicky as betting and stuff like that.

I think one thing I could do to allow folding, would be to make it so that there is an implied bet by the hand that is drawing the least, so the fold criteria are only considered if there is at least one opponent who drew less than you on the previous draw.

The thing I'm a little unsure about is how useful this information is going to be. In a hot/cold sim I'm calculating EV based on how often each hand wins. In a full sim I would be calculating EV by counting how many bets each hand wins or loses. These are two completely different things, and if I just add ability to fold but am not counting bets, you're going to get some weird value in between the two.

In fact maybe the results will be skewed in the wrong direction. Giving the drawing hand the ability to fold will reduce its EV in the hot/cold sim (due to preventing long-shot suckouts that would never happen in reality), but in a real sim surely this should increase its EV, due to the fact that it is now going to save bets?

Isn't that accomplished simply by using a fold criterium?

I think it is going to be somewhat useful for shortening calculation times in problems where you want to try for a suckout. Very far from actually solving those problems, but shortening the time to make some estimates.

That depends on what you compare it to. It certainly loses bets compared to being all-in, but compared to implied odds it gains.

A way to nuance this, and make it simpler to estimate EV, is to include more details than just an equity number.

For instance, a draw that loses 67 times out of 100 might fold 20 times on 2nd street, 30 times on 3d and only 17 times on the river, and then we know to weigh those losses less than draws going to the river and missing. And the same of course goes for wins; the ones caused by folds are less worth than the ones that go to the river.

Another feature in the same vein is when you want to estimate the probability of improving to some threshold *and* losing, when playing draw vs draw. For instance, if I have 987 vs 432 in badugi and want to know how likely it is that my opponent fills to better at the same time I do.

By doing this the user gets several different types of wins and losses, instead of just an average, and can then decide for herself how to tune and weigh the results.

I don't think any of these functions are going to do much beside making estimates easier (and less time consuming, which is a big factor if you plan to spend much time analyzing these games), but they should be relatively low-cost to do.

In order to allow folding we FIRST need to allow betting. What I'm suggesting is, if I add some setting for "fold if still drawing 2", it should ONLY apply that setting if at least one opponent drew less than you on the previous draw because at least that way it is likely that an opponent has bet. Obv you aren't folding if your opponent drew more than you and gives you a free card.

It does give separate win/lose/tie % for badugis and 3 card hands, which I think gives you that info, i.e. you can see how often you make a badugi and still lose, or how often you win by just improving to a better 3 card hand, etc.

Ah, I am suggesting to simply leave all plausibility to the user. Just have an option that can be activated if the user wants to look at a situation where she thinks someone will fold if they don't improve.

Sure, it can be used to look at stupid, or misleading, scenarios, but those are fairly easy to avoid. More importantly, I doubt it is worth the extra clomplexity to have the program evaluate whether a scenario is realistic or not since that is a pretty difficult question.

Oops.

Yes, that is what I was thinking, and if possible even more detailed. For instance, listing the possible combinations of when the players improve/make their hands and what % they are. In the example below, you'd play the hand very differently if they both make a badugi after the second draw compared to if only one did.

This will strongly imply what betting lines are likely to be used, and you can use that to estimate the value of different strategies.

I really don't want to badger you into working, so keep in mind that these only are ideas. Just ignore them if you aren't interested.

Lol don't worry, all feedback is appreciated, though I can't say for sure what I'm actually going to do (and when), it's useful to know what sort of things people want from it.

ok i've read your thinking, and yes there are some problems like if both players are drawing 2 on 3rd draw, but like other guy said i think the user should figure this out. maybe have options (file,edit,options,help) setting where you can click advanced settings or something.

anyway, this is my thinking and why i think it would be cool.

if you implement the folding options, you can have random hands like pokerstove does.

for example, to sim button vs. blinds, suppose you input your hand as say 2368(K). now for next two hands make them (random).

now you can sim, assuming you have folding options on all 3 draws, like

1st draw fold if drawing 3
2nd draw fold if drawing 3
3rd draw fold if drawing 2

and you can tweak it to maybe make one player more tight

1st draw fold if drawing 3
2nd draw fold if drawing 2
3rd draw fold if drawing 2

and stuff.

now, i think one other thing is necessary. my understanding is that you keep all the wheel cards and all cards under the "keep" setting.

But, I think to implement a random hand deal, you should have a "need" setting. this combined with the folding options can let you clearly define a hand.

for example, random hand, need = 2, 1st draw fold if drawing 3, and this means 2ww, right? well 2ww or 2www or 2wwww.

wow another can of worms though, guys can only start with wheel cards. i mean you can keep 6, need 2, but then a hand like 8632, which is a very good hand, will be treated like 236 discard two.

well, for non wheel hands i need to think about it maybe seems like the keep and need don't suffice for describing hands you want like 8632. maybe if you made random hands you could just list the hands you would start with, and the program would pick out the first one that matched so if the random hand was 23678 and the list was

...
87632
...
7632
...
732
...
632
...

then it would take them in order. and the user would have to input the list and the order.

anyway these are my thoughts, great program, thanks.

Can this be configured to calculate for Padooki instead of Badugi?

I think the differences between the 2 games are minor, and come down to what constitutes a flush and how flushes and pairs are scored, but apart from that they're the same game. I've found Padooki seems to get played more often over here in the UK than Badugi does, so this calc would be of interest to me if there were some way of maybe changing a couple of setting manually.

Thanks!

Do you know if it is spread in any card rooms in London?

I found this which says the hands are ranked as follows:
At first I thought this meant it's basically A-5 triple draw with 4 cards, with the exception that badugis beat non-badugis. But I'm not sure if point 3 means that paired cards don't play - but how can it be your "4 lowest cards" if you aren't playing 4 cards?

Does A336 beat A455 (both non-badugi), i.e. are they played as 3 card hands A36 vs A45, or does the pair rank matter?


*edit* Looking at it again, I have to wonder if the guy who wrote that article just didn't understand the rules or had someone explain them to him incorrectly, this disclaimer is at the end:
Is there any evidence other than this article, for the existence of this alternative "padooki" game?

To be honest I've only played Badugi a couple of times (on Doylesroom I think long ago), so can't remember the exact differences between the 2 games, but I seem to remember being surprised that my lesser hand got beaten by someone else's lesser hand when I was expecting to win the pot.

Anyway, to try to make things clear, this is how Padooki (sometimes spelled Padouki) is played, you will then be able to tell for yourself how it differs from Badugi.


- It is a 4 card drawing game with 3 draws
- It tends to be played pot or no limit
- a Padooki is 4 cards all of different suits. The lower your 4 cards, the better
- The hand strength is calculated the same way as a low hand in split pot games - ie it's the lowest high card that wins (so 9876 would beat T32A)
- aces are always low
- straights don't count but they aren't illegal, thus A234 Padooki is the stone nuts
- All 4 cards ALWAYS play for your final hand. If you do not have a Padooki, then you have one of two other holdings - either a flush (where at least two of your cards are of the same suit) or a pair (which ought to be self explanatory). I suppose trips and quads are possible, but you'd have to suck really badly to ever get to the third drawing round and still be drawing 2 or 3 cards
- Flushes beat pairs. When it's flush vs flush - the lowest high card acts as decider. When it's pair vs pair, I *think* (this happens so rarely) that it's the lowest pair that wins the hand. If I was dealt a strong flush (like a 4 or a 5) I might well stand pat a jam the pot on every round. I'm not sure if that'd be a viable tactic in Badugi or not.


I'm pretty sure that's everything. From what I can remember it's almost identical to Badugi, but I seem to remember in Badugi you can have 3 card and 2 card hands. My memory may be letting me down though.






As far as I know, the game isn't officially spread in any card room I've been in. However, I have seen it in a handful of mixed games - mainly in the SouthWest (but that's probably just because that's where I come from, and I've played down here more than anywhere else). I imagine if it's played *anywhere* in London, one of the higher limit dealer's choice games at the Vic would be my best guess.

Edit: After Googling "Padouki" I see Barney Boatman making a reference to it on the Hendon Mob website. This suggests a couple of things - 1) that it's probably played in card rooms around the London area, and 2) there's some easy money to be made if you can find what game he's in

Wikipedia to the rescue. Ok so according to Wikipedia:

This is where Badugi and Padooki differ.

To give a couple of examples where a player of one game may get tripped up when playing the other:


Player A: As 3c 4h 9h
Player B: 2c 4s 4d 7s

In both games player A wins, but for different reasons. In Badugi, Player A holds an A34 3-card badugi, which beats player B's 247 3-card badugi. In Padooki, player A wins with his 7 high flush, beating player B's pair.

Player A: Ac 2h 3s Kh
Player B: Ac 4h 5h 9d

In Badugi, player A wins with an A23 badugi, beating player B's A45. In Padooki, however, player B's 9 high flush beats player A's K high flush.

finally

Player A: Td Jh Qh Ks
Player B: Ad 2h 2c 3s

In Badugi, player B wins with an A23. In Padooki, player A wins with a K high flush.

I'm trying to remember whether or not 3 and 4 card flushes are legal in Padooki. To be honest, given that it's such bad strategy to be drawing more than one card ever anyway, I don't really see those situations crop up except in the super loose and wild games.

Anyway, hope this helps

TL;DR yet, but I am gonna read it later. Thanx for this.

I have been missing badugi soo much, I was playing play money at Carbon.

BIC NEWS

Pokerstars is adding Badugi.

This is probably worthy of it's own thread, IMO. But I am not a regular in this forum, so I will leave that up to a mod, or other poster.

Please don't be a photoshop...

Please don't be a photoshop...

That's awesome, I'd better get some practice in so I can be ready for the first month gold rush

I remember when tribeca added it. It was absolutely nuts the first couple months.