Tuesday, April 19, 2005

Shake your Money, Maker

YTD: +$41871.53

We have looked at previously how a lot of people are overplaying draws online. In fact I have passed *a lot* of nut flush draws recently because I thought the situation wasn't right. Let's now turn this on its head and look at how drawing hands can be a money maker in the right situation.

Of course the best way to work out the profitability of these things is to run through the maths yourself. Back in the day, I actually, geekily, used to do this with pen and paper and then on spreadsheets. Now there are a whole range of free simulators out there on the Net for you to play with. However, for the less mathematically inclined, I’ve put together a simple check list that can guide your play:

1.Is the pot big?
Now you don’t necessarily want it to be so huge that there can be no real action left, because then you will have probably made a different kind of error in that you will have put too much money in preflop in relation to the stack sizes in play.

2.Is the pot multiway?
Although that normally means that some of your drawing outs are in your foe’s hands, this is compensated by the much greater money you can win.

3.Are you drawing to the nuts?

If you are lumping it in on a draw, it normally makes sense to make sure that if you hit, you win. Especially against more than one opponent. Good nut draws are of course our old friend the nut flush draw, but are also straight draws where you have more than 8 cards to hit, often called a “wrap”.

4.Have you got another draw?
If you have another draw as well as your main draw this can make a huge difference. Even if it is as little as a middle pin straight draw or two pair.

If you can answer yes to all these points, it is almost certainly mathematically correct to play your draw and probably play it fast.

Here’s an example from my own play. I called a preflop raise with A K J 7 double suited from a loose player and we went to the flop three-way. We all had healthy stacks of $600 or so. The flop came K J 9, giving me two of my suit and also top two pair. The initial raiser bet and I raised him, got reraised by the player behind me and we all got our money in on the flop. Of course they both had the QT and there was much grumbling when the river gave me the nut flush. But if you look at the maths of the situation, with three way action, I was actually a monster favourite, having over 50% chance of winning the pot. As the breakeven point in this situation is 33% this is a massive edge

It may have looked like gambling, but in fact it was a sure money maker.

13 comments:

Anonymous said...

From Peter B

Hi Dave: Jeez, with top two-pair and a draw to the nut flush, I would have been happy to go all-in against ONE opponent if I reckoned he had the dry nuts. Against two, I'd have happily put my grand in alongside yours.

But what if you had (say) top and bottom pair plus a draw to the king flush? Or if you had bottom set and a gutshot? These are the ones that I find tough. Presumably your advice would be to either walk away or lump it in (no point in limping) but, unless I am up against a known rock who wouldn't be playing the hand without a made straight and redraws to the nuts, I really don't know what to do with such a marginal.

I used to have this problem in limit (particularly pre-flop) but now even the marginals are automatic decisions. I probably get one hand a week where I have to have a little think about the maths of a situation (kind of like A6 off in seat five of 10, no previous callers, a loosey in the small blind and a rock in the big blind. Could be a raising hand, could be a folder.)

More PLO hands on the www.livejournal.com/users/_pjb_ fiasco. It's fun fun fun at Borks Towers.

steve said...

Peter,

Can't seem to look at your journal, should I do something to that address?

Secondly, you said this re LHE:

'I used to have this problem in limit (particularly pre-flop) but now even the marginals are automatic decisions. I probably get one hand a week where I have to have a little think about the maths of a situation (kind of like A6 off in seat five of 10, no previous callers, a loosey in the small blind and a rock in the big blind. Could be a raising hand, could be a folder.)'

I found this surprising and interesting and got me thinking a little.

Now, I would say that pre-flop decision-making was one of the weakest areas of the game, particularly in or around the blinds (still is, but I’ve shored it up a bit). Nothing surprising there: it's hard to match up your hand against the range of hands on show in late positions + there are other murkier benefits to be considered there.

Now this is an area I've worked on and improved upon, but to be honest there are still areas where I feel rather clueless. Particularly as many hands are hands that have a 'mix of pre-flop strategies': it certainly isn’t deterministic. On top of that there is your 'local image' to consider (locally tight, locally weak etc). But even decisions that are isolated from these factors are still very unclear (PT sample size isn’t big enough, imo, but certainly provides clues).

I think it is desirable and partly inevitable that we develop a template or a strategy that we’re comfortable, feels right and we can consistently apply. There is plenty of benefit in this, however, it can give a false sense of security. Decisions that appear marginal on our strategy aren’t necessarily marginal at all (against some ‘true strategy’, unique to our game). In time we develop a skilful ability to make marginal decisions based on or around our strategy. This gives us confidence.

But there is a danger that we spend our time perfecting our decision-making within the strategy rather than developing the strategy itself. I’m searching for a line, but I can’t quite find it (I’m sure Dave will come up with one), but there is clearly difference between being good at making, perceived, marginal decisions and making marginally-right decisions (and that ain’t the line!)

As an analogy consider Blackjack. There are a raft of strategies all giving different results and strategies and crucially, different views on different situations. Now a player may use a system and discover a marginal situation, say 16 v 10. The player has practised this system and manages to quickly, accurately assess that the count is just greater than 0 and stays on 16. Now he can be confident in his game and his ability to make this and other marginal decisions based on this strategy. But against another count/strategy, more or less sophisticated, (or against the ‘truth’ in some sense) this situation might not be marginal at all: his confident decision may in fact be wrong. By focussing on his ability to apply the strategy rather than develop it, he has traded efficiency for effectiveness. Another player, less efficient, but with a better strategy may be more effective. The first player will probably feel happier with his game than the second player who is making more 'mistakes'.

Peter I’m certainly not directing this at you (I know little about your game/results – you may be both efficient and effective), but your statement made me think about this. It is a complacency that we must all be on our guard against – particularly myself, recently.

Now there’s another article I might never write, but I hope it made some sense.

Dave: A little pop at you for the 20:20 thing, what do you think your chances were on the balance of probabilities – do you do this on hand reviews? I realise it must be a bit tough.

Also is there anyway of cutting and pasting or drawing things on Blogger, say, 3 ovals and three arrows?

chaos

Anonymous said...

From Peter B

Chaos.

www.livejournal.com/users/_pjb_
(where the "pjb" has an underscore on either side of it) should work.

Or try just www.livejournal.com and search for _pjb_

I take your point re automatic decisions within a strategy, rather than developing the strategy itself. Indeed, I cover just this point in a limit hand on my blog, where I "go limp" with Aces post-flop against one (known aggressive) opponent. This works at $3-$6, but usually not at $2-$4.

The reason that most of my actions pre-flop are now fairly automatic is that I play so many limit hands a month. I've tried various limping/raising/ aggressive/passive strategies, and pokertracker tends to tell me which ones work and which ones do not.

"Table image" and who your opponents are becomes more of a factor at the higher levels. Usually at 2-4 and 3-6 I just know who I have played against A LOT, and if necessary I adjust my play. But there is a danger of getting into a "too clever for yourself" mode at these lower limits. You multi-table. You play ABC, you play a strategy that is one stakes level "higher" if you know that your opponent is a winner at this level. If your opponent is a complete fish, then you can play one level lower.

All of this is by now pretty automatic. So, no, it doesn't mean that I will play Ace-Nine off in late the same way every time if there has been one limper. What it does mean is that I know when it is right to raise and when it is right to pass. And I know it without thinking about it.

That doesn't mean that post-flop, turn and river situatiuons don't occur where I need to think quite hard. This is because these situations necessarily occur less often (since I only see about 20% of all flops).

Pete

Big Dave D said...

Pete,

If you routinely jam headsup in that coup, then you are losing money.

Chaos,

The difference in effectiveness vs efficieny is also dependent on how high you are playing and the types of games you are in. The bulj of winning players are multitabling 15-30 or less. Here it is probably clear that efficiency rules, as the addition of extra hands/tables is "probably" more +EV than playing any particular hand more effectively, within boundary limits.

As to the 20:20 thing, after every session I review the big hands I lost and make a noddy view of whether I played them good or bad. If it is really a challanging decision, and they do occur, then I do a proper range of hands EV analysis. And some interesting things do occassionaly pop out which I will share here.

Maybe :)

dd

SimonG. said...

I hope Dave doesn't mind us talking amongst ourselves on his blog! I'm sure he won't becuase I suspect he is tired of all this PLO action and yearns for some low limit holdem 'action'....

Pete, I think you may be "too clever" for the level you are playing. I can't argue with what pokertracker is telling you and you of course should be happy to follow what it is pointing out. But when it comes down to working out if you need A7 or A6 as your minimum based on the blinds' propensity to fold, I believe you are putting in waaaay too much effort!
I agree that after playing limit holdem day in, day out for a long time (as I pretty much have also) most hands become automatic. The 80%/20% Pareto principle (if that's what it is called!) doesn't apply online because people don't pay attention, but there's no harm with a 90/10 rule. One of the benefits of playing this way is that you don't have to worry about making a decision correctly to within 3% all the time. even if you make a decision on the wrong side of 47/53, your decision may still be fine if you cover the missing 3% from advertising value, mix-up value, or any other reason. i.e. your decision can often turn out ok (allbeit marginal) but just not for the reason you originally thought?
Although I like to look at the stats, in your A/6 example I would be far more likely to base my decision on what has recently happened rather than my notes on that player over the last year. If I had been on the rob recently or shown down a few ropey hands, I may pass A6 if it is approximately marginal. Flipside, if I had shown down premium recently or been quite quiet for a while (yeah right!) I may well play the hand for a raise. I would certainly place more relevance on "feel" here than the numbers.

As always, many different ways to skin that cat - not questioning your ability to get the job done, just suggesting you are using a sledgehammer to crack a walnut! :)

Big Dave D said...

The more debate the better, Simon.

You are basically stating Chaos's point more colloquially. The problem is, say you werent watching the last few hands?

But you would say, why are you not watching? Because it is more profitable to be playing more hands than to be playing better with the hands you play.

gl

dd

steve said...

Pete,

I’m very doubtful that PT can inform that well. Take AA UTG, UTG+1. What sample size would you need for a comparison study? A few thousand would be a minimum to usefully inform, but even that is doubtful (if it really is marginal). That would translate to couple of million total hands. Add to that any further filtering you might desire on top of position on top of that e.g pot sizes of about 8 big bets and you are in intractable country w.r.t hands.

By a defined strategy I don’t mean always passing A9o to a limper, it might be, but it can be defined to pass it against a solid limper and raise against a loose player. But it is tough to elicit information to determine that raising a loose limper with A9o is marginal a decision and tougher still to know which way to turn from hand tracking software. It might be possible to group together certain decisions and make some inferences, but so refined are some marginals that not only is a large sample size required, but the facility to filter them out of the database might be beyond the software. Add to that your game is likely to change over such a long sample period, you may, for example limp better with AA, after 2 million hands than before. Consequently any analysis could misinform.

The trick to discovering an effective strategy maybe to use PT in another way.

Re being too clever – I agree, it is certainly a problem with mulit-tabling, often you don’t know whether you are being too clever or not clever enough.

Dave, I agree except I think that what you are saying leads to greater effectiveness but less efficiency: i.e. you are less efficient in your decision making but more effective because you make in up in volume. But I’m not too sure on the language, I can see how you can say it is more efficient. It probably depends on your sphere of influence!
(but I must give deference to you business/economics guys)

In this context I was thinking of efficiency as a measure of how well you employ your strategy and overall-effectiveness is how much you value that efficiency based on the strategy you choose. I’m sure many players choose to play a more effective strategy but play it inefficiently and are better off choosing a less effective strategy you can play more efficiently. Maybe ‘productivity’ should come into this somewhere – some function of player-efficiency and effectiveness of the chosen strategy.

Just looked at Simon's can't disagree with the points made

*by effectiveness w.r.t. strategy I mean how useful or product it if it is played efficiently. BJ is a good example here, because you can develop very sophisticated counts compared to simple +-1 strategies, but they are very hard to employ.

steve said...

btw no one answered my lines and ovals problem!

Anonymous said...

From Peter B

Chaso/Simon G: My limit comments were more throwaway lines (I was, after all, talking about PLO at the time) than a deeply thought-out statement of policy. Even the hand selection (A6, or whatever it was) was dragged up with only a few milliseconds thought. Of course, I also use things like "whhat is my current table image" and the like, and I probably use more "feel" than I am aware of, even pre-flop. Then again, if I am using more feel than I am aware of, then the action is automatic!

On the sample size in PT, what I do is group hands into patterns. This generates a good analysis with, say, 30,000 hands as a sample size.

BTW, I like SimonG's point about it not really mattering if you make the occasional wrong 47/53 decision, but in the long run, if you make a habit of making those decisions, it costs you money. Feeney makes just this observation, that a lot of very good players make some losing plays (not big losing plays, but marginal losing plays) and they make them consistently. Why, Feeney asks, do they choose to throw away money? The answer, of course, might be that these marginal losing plays more than pay for themselves by changing the way other people react to you. But it's only a hypothesis that it changes the way other people react to you, not a certainty. Fox's line, which is to always play properly, but then to lie, strikes me as a better way to go about it.

Pete

SimonG. said...

Glad to see some good discussion.
Re-reading Chaos' comments I now see I have a more colloquial version of the same - I think I will always have a more colloquial version though! :)
The point about 47/53 was taken as I meant (i.e. an occasional mistake is not too bad, particularly if it turns out not to be a mistake for a different reason)

I of course wouldn't recommend making decisions 3% worse than your opponent in the long run.

Another example of being happy to take the 47/53 comes from big bet poker where a quick decision is more important than the marginal wrong decision when you don't want your opponent to know you have just agonised over a tough decision.
Although I know little about football, I liken it to a referee blowing for offside. When it takes 5 replays for the pundits to work out if he was right or wrong, neither decision can be that bad - it is more important that he makes a quick decision and goes with it.

Of course, in online poker that doesn't hold up as the delay may have been while he referred to pokertracker(!) but it may because of his conncection speed, multitabling or multitasking.

steve said...

Peter,

I think it was the one preflop decision a week , that caught my eye rather than any specific dilemma. This is certainly something I din't relate to.

Re: the sample size: if your sample size is 30k then you must have a mother of a database.

Simon's: 53:47 ~ Unfortunately, I'm sure I have lots of these in my game, but I'd happily trade 10 53:47's for 1 80:20.

Example: A little while ago, in a tight ring game, in which my local image was fairly tight, I raised utg with 10-60. I was reraised, by a solid player. Anyway I hit a 10 and he called me down with AK.
Response:

'wtf'
'oh you must be one of the greats'
What odds that this play elicited a note or two to my name? One decision and an opinion was formed, for all he knew I could have hit the wrong button!

Simon said

'Another example of being happy to take the 47/53 comes from big bet poker where a quick decision is more important than the marginal wrong decision when you don't want your opponent to know you have just agonised over a tough decision.'

Good observation Simon: I'm not sure whether you are saying 'you are happy' in the sense that you should be, or simply that you/we are. It has irritated me no end that 'I am happy to do so' because I am pathetically trading EV for some ego-stroking. The value of which dissipates within seconds/minutes of the completion of the hand.

And still no answer!

chaos

Big Dave D said...

Chaos,

I think the answer is no, unless u bring it in as a picture. Don't ask me how to do that.

I get your efficency effectiveness point now...as always you are looking at it a level deeper than me :(

My holdem strategy probably filled your latter requirement, it was efficient more than it was effective. I basically had 3 decision branches - I had raised; I had called in a value pot with position; I had defended my blind. That was basically it. With PT for an insight into players in the hand.

Simon,

The reason the dwell up was such a good tell back in the day was that moodies were strictly forbidden. In today's world Im not sure how true that would be.

gl

dd

Big Dave D said...

Hey Joe

Good to see you here and good to see you still in the tuffist game in town. Although u wouldnt think it by this hand :(

I think your line is good, but so is going allin on the flop. The times that a player passes for the exta 500 on the flop (small) are balanced out by the number of times u get someone to pass on the turn (also small)

gl and keep posting!

dd