Irrationality

[peterbirks]

The markets can remain irrational longer than you can remain solvent, said JMK, a man remembered mainly for Keynesian economics but who perhaps should be more remembered for his success in stopping the Truman administration causing starvation in the streets of London (through ignorance rather than maliciousness, btw).

This week the thought of LTCM (and Keynes) came to mind when I saw that it was now marginally more expensive to buy municipal bonds insured by MBIA or Ambac than it was to buy bonds with no insurance at all. Now, it doesn't matter how likely you think MBIA or Ambac are to fail, that likelihood, unless you are Patrick HanrahanOHanrahan, cannot exceed 100%. Therefore the insurance offered by MBIA and Ambac cannot have a negative value. Therefore the pricing is irrational and it will, eventually, come back into sync. This would be the kind of discrepancy that LTCM would have pissed on in the years before it went mad and stopped realizing that the medium-term is slightly further away than just after lunch.

The study of irrationality is hip. For so long a fundamental of economics has been that people act rationally, even though common sense tells us that this is not the case. Although I know that I should read more poker books, I am irresistibly drawn to books which I think tell me more about the meta-world of poker players and what makes them tick.

Sway looks to be the latest "must-have" in this field, following the fad of Nudge and Predictably Irrational.

In poker, it's a great strength to know how people's irrationality can be exaggerated. Why, for example, do we find ourselves playing more cautiously if we are near to level as the end of a long cash game approaches? If we are a long way up or a long way down, we are far more likely to make gambling plays. To what extent does that fact that telling someone that it's their looseness that helps them win (when in fact it's been their tight-aggressive play that has won them money) get them to play more loosely? (If you tell someone that they are a bit deaf then their subsequent score in a hearing test will probably drop.)

To what extent to skilled "chatty" live players (see Negreanu, Channing) exploit this kind of stuff?

If you pass a guy on the street looking up, how likely are you to look up?

++++++++++++

I've been five-tabling this week, partly because my energy level in the gym has convinced me that I'm in a bit of an up-phase of the good ol' bi-polarity. I've done this to stop my normal reaction, which is to shift rapidly up in stakes. This led me to consider other matters of meta-rules that matter far more, I think, than individual marginal decisions.

1) I can manage just about 70 minutes of five-tabling before my attention starts to wander. When my attention does start to wander, I have a habit of calling preflop with hands where I should really fold or raise. The important question I have to ask myself pre-flop (about 40 times an hour out of the 350 or so hands played per hour) is "what is my game-plan with this hand?" If the only answer is "well, let's see what happens on the flop", then a call is almost certainly wrong. Indeed, calls are very rarely right (for my style of play in the type of game I find myself). An implicit assumption in many poker books appears to be that there will be a drip-drip loss of funds (because you see speculative flops cheaplpy and then fold because you miss) compensated for by occasional big wins (where you hit your hand and get paid off by a serial non-believer). Perhaps there are many games like this, but if you find the big wins disappearing into the aether because you don't get paid off, what you see is a drip-drip loss that is NOT compensated for by occasional big wins.

I would say that the majority of tightish players (say, those with a 17%/4% range of stats) lose money gradually because they are not getting paid off. Their opponents have Pokertracker too. They can see the numbers. These guys' bet sizes are rather unimaginative, which makes reading their hand strength even easier. Only the (very few) absolute fish pay them off. Result, a gradual loss of cash which the player blames on poor cards.

Meanwhile there's another kind of player (say, 16%/16%) who also gets into trouble. Once again the opponents have Pokertracker, once again they can read the stats. This guy puts in his raise in the cut-off and can't figure out why either the button or one of the blinds always seems to "find a hand" with which to reraise. The net result here is again a gradual loss -- this time caused by folds preflop to a reraise.

The winners (and I emphasize here that I'm talking about these afternoon games where the majority of the players are multi-tablers of the grinding/nitty variety) have a wide-range of VPIP$, from as low as 9% to as high as 25%. But the proportion of raises remains roughly constant -- declining slightly as you reach the higher VPIP%. Say, from 9%/6% to 25%/17%.

++++++++++++++

I've slightly adapted my style in the past few weeks, playing marginally tighter and reducing my raise level to 3x from no matter where (4x with one limper). Occasionally I vary this depending on the stack size of people behind me and on how loose/tight aggressive/passive they are. I'm juggling with the idea of reducing my bet size on the flop with a view to taking it down on the turn more often. I've discovered that I make more when I get called pre-flop than I do when I am not called (yes, one would expect this, I know), so, quite possibly, I'll make more if I am called on the flop than if I elicit a fold from my continuation bet. One argument here is to widen my range (I know this one), but I like the style of making it look to opponent that he is getting implied odds when he is not — i.e., offering a cheaper price that is still too expensive).

In effect, I'm trying to get opponent to make the mistake I find myself making when I begin to lose concentration. Calling to see another card with no real hand-plan, but just "because it's cheap".

___________________

Comments

It all adds up.

[(Anonymous)]

Glad to see I'm not the only person who remembers the great journalist Peter (not Patrick) Hanraha O'Hanrahan.

You are probably thinking of this report -

http://www.youtube.com/watch?v=3RpFPCDgeI4

though I think this was his finest hour -

http://www.youtube.com/watch?v=7Bq_dkPkQUU&feature=related

I recently discovered this radio report he did on September 11th 2001 -

http://www.youtube.com/watch?v=1SPWgodul_E&feature=related

DY

Re: It all adds up.

[peterbirks.livejournal.com]

Well, it's Patrick Marber, isn't it? Hence the hurried error. But, really David, you are going to have to stop this YouTube trawling. It's addictive, you know. Next thing you know you'll be looking up classic Fist of Fun episodes.

PJ

Re: It all adds up.

[peterbirks.livejournal.com]

Actually, that wasn't the clip I was thinking of. I had in mind some story where he referred to 115% (or 110%) and Chris Morris asked HanraOHanrahan how that was possible.

Then again, perhaps my memories playing tricks.

PJ

Things are looking up (an elephant's bum)

[real-aardvark.livejournal.com]

"If you pass a guy on the street looking up, how likely are you to look up?"

Since I am supposedly already looking up, the chances are quite high. I have no idea whether I would flip directions if the other guy was also looking up, however.

... but, I accept your premise. I don't, however, think that the madness of crowds can be transposed to individuals (otherwise we'd all be drooling in Bedlam, and then who would be outside to pay threepence for the privilege of watching us? Perhaps in a PKD universe this is already true). As a misanthrope, I pay very little attention indeed to what other people do, unless they start defending the nutritional benefits of lettuce, in which case I apparently push them through the opposite wall.

You're remembering JMK for all the wrong things. I remember him as the Old Boy who single-handedly built up the funding for Kings College Cambridge by astute investments (presumably in third-hand beauty contests and the like -- I'm a Historian, so the details don't make no nevermind) until it was the second richest college in Cambridge. (After Trinity College, which obviously had a lot of apples left over by Newton and stuff.)

Now Kings College is technically bankrupt, or so I was told when I stayed there a couple of years ago. This has the obvious benefit that they have to whore around by pitching summer programs (sic) to gorgeous pouting blonde co-eds from Ivy League schools (sic transit et) with rich daddies, but it does lead me to muse on the decline of Keynesianism and the consequent damage by Keynes' other half (Milton, obviously) to places that I love.

I think the concept of a negative price margin on an insurance bond is superb. I will inform the current bursar of Kings post haste.

It's a bit sad that, purely as an afterthought, Keynes managed to alleviate the plight of the starving ragamuffins of London in the '50s. I'd never recommend genocide or indeed darwinian selection to anybody, but that would have worked for me.

In unrelated news: Tiger Woods seems to have nothing much to do for the rest of the year. Can we assume that his next child is going to be called "Bretton?" The sums of money involved are probably about the same.

Maybe a little rational

[(Anonymous)]

I am not sure this explains every such seeming irrationality, but keep in mind wrapped bond provisions typically limit your ability to accelerate against an issuer default if the wrap provider is making each payment for them.

So, imagine two low coupon bonds issued by Deadbeat city, one wrapped by Shaky assurance co and the other one not. Deadbeat declares a moratorium on its debt, an event of default under both bonds.

With the unwrapped bond, you accelerate, get in line with your claim, chase their assets and collect what you can in a few years, all without having to deal with Shaky.

With the wrapped bond, consider the risk that upon a Deadbeat default, Shaky continues to make interest payments trying to delay being on the hook for the principal all at once, so you can't accelerate. If by the time Shaky goes tits up, the value of the interim coupons paid by Shaky doesn't exceed the drop you anticipate in the ultimate delayed principal recovery scenario (particularly true if Shaky is really Shaky) as well as your additional market exposure to Deadbeat while waiting, you might have basis for preferring the unwrapped bonds.

Just two cents, now worth, as you well know, only one pence.

SxLvr

Re: Maybe a little rational

[peterbirks.livejournal.com]

Nice point!

I'm not sure of the technicalities of this "accelerate", but I can see where you are coming from. There is a kind of parallel here with asbestos-related diseases, where the really sick (except they haven't become ill yet) are going to find themselves chasing an empty pot, because loads of people with pleural plaques will have been compensated for worrying that they might become sick. In 15 years time, the people with pleural plaques will in the main, remain resolutely healthy (and well-compensated), whereas others who did not develop these "pre-indicators" could suddenly come down with a much more serious condition (say, mesothelemia -- a very cruel cancer that creates something akin to an ice-hockey puck in your chest) who have shown no pre-indication.

But, I digress.

So, you have a bond muni insurance policy. Deadbeat City goes bust. It tells you to piss off "because you are covered". Other people get about 20 cents on the dollar. Then Shaky Insurance goes into default, and pays out only 15 cents on the dollar (and does so several years subsequently). That would theoretically explain a negative value for the insurance on the bond.

Unfortunately, nice though that line is, I don't think it has any bearing in practice. If Deadbeat City goes bust AND Shaky Insurance goes bust, you might have to wait longer on Shaky Insurance, but you would almost certainly end up getting significantly more cents on the dollar.

PJ

[(Anonymous)]

Peter,

do you use PokerTracker or PokerOffice? Is there a choice of versions?

DY

[peterbirks.livejournal.com]

I use Pokertracker 3 now David, where I can. It's not yet available for some of the sites, so I have to use PT2 there. PT3 has PAHud integrated, which makes like somewhat easier. However, I'm still importing all the hands into PT2 every so often, just because I'm more used to reading the stats of that page.

PT3 is horrifically powerfuil if you are a programming geek because you can write your own stats formulae. Unfortunately I don't know the relevant language and I don't seem to be devoting the time to learning it (other things in life get in the way).

So I'm waiting for some customised stats that I want (or can at least adapt) to appear on the PT forum.

It says something of the general quality of the players in the world that stats seen as "useful" are "number of times flopped set". Well, screw me, I know roughly how many times I'm going to flop a set in 100,000 hands, thanks very much. What I'd like to know is how often I'm floated by the button when I raise to 3x compared to when I raise to 4x, given full stacks (and all the varuious permutations therein). How often does my CB OOP take it down uincontested compared with a CB to a blind? Questions like that.

These things seem harder to write, but they'd be darned useful to know.

PJ

Garbo Geeks!

[real-aardvark.livejournal.com]

Well ... That's interesting. Winding down on a Friday evening without benefit of wine, women or song -- as one does -- I googled PT3 and language/plugin support. No joy on the first couple of pages. So I went to the site. No joy.

Are you sure that PT has actual language support?

I've actually downloaded the trial version (please bring back the wine, women and song of yesteryear), and there's no evidence of this. There is, however, evidence that it stores everything in a dirty great database -- PostGRES if you go for the obvious, and God knows what if you don't.

This is not necessarily a bad thing. If the schema is intelligible (and I obviously don't know, because I'd have to be insane to play enough on-line poker to check this), it should be a relatively trivial task to suck the data out of the database and play with it as you will.

"How often does my CB OOP take it down uincontested compared with a CB to a blind?"

If I'm correct in my assumptions, that would be at most a two-hour SQL programming job in something like Perl.

Is it worth twenty quid to you, guv?

Re: Garbo Geeks!

[peterbirks.livejournal.com]

If you click on the "documentation" tab on the home page, you get taken to a whole raft of guides. The "statistical reference guide" lists the relevant commands available and the "custom statistics guide" is meant to help you implement them.

It's possible to have two databases (e.g., one for hands that you play and one for datamining), although they can't be open simultaneously. I assume that I can access the same database from my laptop, but I haven't set up PT3 on the laptop yet.

When Pokertracker started it used Access as the database schema, not imagining that people would be building up 1m hand databases that basically slowed the whole thing to a crawl. PT 2 offered a PostGRES option about three years ago, but I never took them up on it, mainly because I had no idea how it worked and I wasn't interested in finding out. What I had worked, and that was fine by me.

What I really need is to sit down behind someone who has set up custom stats and then a custom report (the latter seems to me to be easier than the former) so that I can see how it's done. A good old "look and do" way of learning. The comments by the techincal expert on the forum are worse than useless. He may know his stuff, but he's hopeless at communicating how to do it.

And, no, it isn't worth twenty quid to me.

PJ

[(Anonymous)]

'To what extent does that fact that telling someone that it's their looseness that helps them win (when in fact it's been their tight-aggressive play that has won them money) get them to play more loosely? (If you tell someone that they are a bit deaf then their subsequent score in a hearing test will probably drop.)'

It's an interesting subject, I think this sounds somewhat akin to the stereotype threat. Although I can see it having the reverse effect in many cases. But I suppose if someone is exactly where you don't want them to be then it is advantageous just to move them off where they are: looser or tighter. A bit like our hand-showing discussion we had way back. At times it should enable you to move your opponent off a place you'd rather him not be - on the money. In general, I'm not all that convinced about it, though, as it creates agendas. But I wrote it here, some time back.

http://pokerdynamics.blogspot.com/2006/06/to-show-or-not-to-show-ii.html

[(Anonymous)]

~ chaos, obv.