Finally, I have a working C++ code to do the epsilon distance calculation. I must be getting old. It's been a right pain in the arse. I seriously thought it would take me about a week to get the whole thing done. Final error? Writing 1-1 instead of l-1 in a vital line. So hard to spot. 1000+ lines of very repetitive C++. Fml.
Really good new post in the GTORangeBuilder blog on the difference between GTO and unexploitable. I hadn't really thought about things in this way before, but he's absolutely spot on as usual.
Going to put out some more files with your epsilon distance working? Slacker.
I thought the "NOSEBLEED" documentary was pretty good. Unfortunately my french isn't good enough to understand what they were talking about so I had to pay attention and manged to watch the whole thing without getting bored.
Poor Gus Hansen, he cops a bagging.
Originally posted by DonErmoso
JB, how is Mrs.JB doing in this days?
Are still the same fish?
She's alright.
The fish are still the same. The same are still fish. Fish still the same. Are. Fish.
Still.
Originally posted by jules97
Going to put out some more files with your epsilon distance working? Slacker.I thought the "NOSEBLEED" documentary was pretty good. Unfortunately my french isn't good enough to understand what they were talking about so I had to pay attention and manged to watch the whole thing without getting bored.
Poor Gus Hansen, he cops a bagging.
/blockquote>
I managed to get MATLAB to use larger files so I can do bigger trees, but it then crashed. Probably a 64 bit thing, but I need to sort it out.
The conclusion from my epsilon distance code was that all the solutions I already published are probably not exploitable by more than 1.5% of the pot, which is pretty fucking awesome, though I say so myself.
Originally posted by jbpatzer
Originally posted by DonErmoso
JB, how is Mrs.JB doing in this days?
Are still the same fish?She's alright.
The fish are still the same. The same are still fish. Fish still the same. Are. Fish.
Still.
I just realized what you meant.
No.
They died.
Originally posted by jbpatzer
I managed to get MATLAB to use larger files so I can do bigger trees, but it then crashed. Probably a 64 bit thing, but I need to sort it out.The conclusion from my epsilon distance code was that all the solutions I already published are probably not exploitable by more than 1.5% of the pot, which is pretty fucking awesome, though I say so myself.
I haven't written anything for a while, so I thought I should put finger to keyboard. It's the last week of term today and, since the first years have really been grinding me down lately, I'm pretty pleased, although that does signal the imminent arrival of Xmas, which is less pleasing.
I fixed the memory problem with my code, which is good. I really do just need to get fractional preflop ranges and multithreading working now, but I'm not feeling terribly motivated at the moment. I've been talking to Alex Sutherland quite a lot, and he's working on the same problem, but with fictitious play. I don't believe that's going to work, but we'll see. He's posted something http://blog.gtorangebuilder.com/2014/12/gtorb-flop-sneak-preview-3-street.html about this, but it's for tiny ranges. My computer is currently working out the solution for the situation he presents here, http://blog.gtorangebuilder.com/2014/11/improving-your-turn-play-with-gto-gto.html , but I'm doing it from the flop. It's a huge pair of ranges, and it's taking a while (multithreading needed!), but the exploitability range is down to about 5.5% of the pot. The exploitability range is the difference between how much each player can exploit the other, which is cheaper to calculate than the individual exploitabilities, because you don't need to calculate the value of the game. If this were symmetric, that would mean plus or minus 2.75% of the pot. Anyway, I'll leave it running for a couple more days, and I should be able to get the exploitability down further, even though it's clearly already a very good solution. I'll then compare it to Alex's video.
One of the reasons I'm not so motivated by my GTO project at the moment is that nobody apart from me, Jules and Alex really seems terribly interested. Tbh, I find this pretty extraordinary. I can find the GTO solution from the flop on giant game trees with wide ranges ffs! I mean how can that not useful?? This has led me to think I should just play more and code less, and use the results from my code to improve my own play (and Jules's). I'm bankrolled for NL100 if I want to be aggressive, and comfortably bankrolled for NL50, so I'm doing what any sensible nit would do and playing some NL25. I'm running like God, Jesus, Mary, Joseph, the Three Wise Men, the Shepherds and Phil Helmuth all rolled into one at the moment. I'm sure that can continue indefinitely. I may post some graphs and hands this month.
One downside of my suddenly renewed interest in four tabling Zoom and getting 1K hands per hour is that going to play live seems inconceivable to me. If I played a six hour live session that's probably 200 hands at best, which is less than 15 minutes of Zoom. Even though I'm running well, there are plenty of stinking 200 hand stretches in the 5Kish hands I've played this month. I just don't think I can do it unless it's to meet someone and be sociable.
Hopefully I will grow some balls and move to NL50 soon. You never know.......:badteeth:
Yeah, people are idiots. 'why learn to play optimal poker? no one plays optimally, just exploit people.. derp derp'.
I've got a bucks show in vegas in a couple of months. Decided to book a flight a little early and play live poker for a week beforehand. I wonder if I can play live for a whole week without killing myself. Seemed like a good idea at the time. We'll see.
Went to the local casino a couple of weeks ago. No decent games were running, and with rake at like a million dollars a pot I didn't bother. Did see some interesting blackjack strategy though. I swear, no exaggeration, a guy did exactly the same thing as Homer Simpson. The dealer was deadpan too. Like he sees this kind of shit regularly. 
Enjoy your run good, this last week I've been running bad and playing awful. Playing some mtt's on the side to help me grind cash games. I keep getting deep into them and then losing 80% flips... So frustrating.
Check out this masterpiece... I'd provide some excuses for my line, but it would probably only make it worse... I should write a book with that blackjack guy on how not to gamble... 
BTN: 126.4 BB
SB: 121.88 BB
Hero (BB): 170.08 BB
UTG: 143.24 BB
MP: 53.68 BB
CO: 100 BB
SB posts SB 0.4 BB, Hero posts BB 1 BB
Pre Flop: (pot: 1.4 BB) Hero has 9♦ 5♦
fold, MP raises to 2 BB, fold, fold, SB calls 1.6 BB, Hero calls 1 BB
Flop: (6 BB, 3 players) T♠ 8♦ 7♦
SB checks, Hero checks, MP checks
Turn: (6 BB, 3 players) T♥
SB checks, Hero checks, MP bets 2 BB, SB calls 2 BB, Hero calls 2 BB
River: (12 BB, 3 players) 6♦
SB checks, Hero checks, MP checks
SB shows T♣ T♦ (Four of a Kind, Tens)
(Pre 68%, Flop 62%, Turn 98%)
Hero shows 9♦ 5♦ (Straight Flush, Nine High)
(Pre 15%, Flop 37%, Turn 2%)
MP mucks 2♦ 2♣ (Two Pair, Tens and Twos)
(Pre 17%, Flop 1%, Turn 0%)
Hero wins 11.44 BB
At least you know how to scoop up part of a bad beat jackpot. That's the sort of slow playing that makes glaciers look speedy. 
New post about the code at https://www.pokerstrategy.com/forum/thread.php?postid=1971941#post1971941
I promised an update, so here it is. As predicted, after running like god initially, I've run like a big pile of poop for about the last 10 days. Not just below EV, but lots of coolers. That's still 5bb/100 hands although my graph, he go up, up, up, then down, down, down (I'm not falling for that 'a picture paints a thousand words nonsense'). This is, however, standard. I feel like I'm playing OK. I just can't get enough volume in.
In lieu of tilting my bankroll away, I went back to my code and included the effect of the rake. It was fairly straightforward, apart from the exploitability subroutine, which was so painful it made me go back and play some more poker. However, it seems to be working now. I need to redo some calculations with the rake added to see how big the effect is, and I need to update the gui as well. I've set Janda hand 1 going again with Pokerstars NL50 rake (4.5% capped at 5bb). I'll check the results later on, but I can tell you straight away that rake takes about 10% of the initial pot at equilibrium (i.e. about 0.9bb out of 8.5bb in there initially). I don't know whether that goes up or down given the way people actually play, but that's a winrate for Pokerstars of about 60bb = $30 per hour per table. Nice work if you can get it.
In spite of what I've said about the turgid nature of live poker, I'm going to go back to DTD on Saturday to play their lolstakes MTT, because it's so cheap and so soft and such good value. I expect little in return but pain, but I'll have had some fun and got away from the kids for a while.
Was just woken up by another murdered koala. Almost got out in time to stop it, about 30 seconds too late...
Yeah it should be interesting to see how rake changes things. The rake at nl50 and lower is ridiculous.
Not much going on here, had a nice 10 buyin day yesterday after a fairly breakeven month so far. One thing that was slow to hit me after looking at all these GTO solutions is how betsizes don't really matter too much. And how so many regs think about using bet sizes. That 'really good players' use 'creative betsizes' to 'target ranges'. Which is fine against fish, but in reality they're turning their hands terribly face up. I think this has helped me quite a bit since it sunk in.
Just because bet sizes don't seem to matter too much at equilibrium doesn't mean that there aren't exploitative reasons to have different bet sizes, or that there isn't an optimal bet size to maximise ev, but basically I think I agree with you.
Preliminary results suggest that rake doesn't matter too much in determining ranges either, but ofc if 10% of your win rate is getting raked away, you probably have to play tighter preflop.
I'm off to DTD to get pwned by shoals of fishies. 
the straightflush vs quads hand is beautiful
and imagine the sudden switch of emotions the guy with quads must have gone through in the space of about a second, first his failure to get a bet so that he can checkraise his quads, then the relief when he sees that he would have been up against a straightflush
Happy new year!
I assume you shipped the live donkament. Then degened your winning away instantly at some roulette or something! 
I didn't bother checking out your file of the Q82 hand, I figured it's close enough to the GTROB one, but, well, the GTROB solution is a lot nicer on the eyes... so I just looked at that one ♦
Did you ever get out the one with rake? Any plans to pump out some more files? I'd really like to see something with a mono flop.
Well, it's been a while since I posted. I'll start with Jules and then move on to my Xmas presents, as is traditional at this time of year.
Originally posted by jules97
Happy new year!
I assume you shipped the live donkament. Then degened your winning away instantly at some roulette or something!
TT<QJo just before the bubble. Pretty standard. I'm seriously thinking of finding a pub to play in instead.
I didn't bother checking out your file of the Q82 hand, I figured it's close enough to the GTROB one, but, well, the GTROB solution is a lot nicer on the eyes... so I just looked at that one
Yes, but mine starts from the flop, not the turn. However, you do raise an interesting point. The GTORB interface is based on the game tree with pop ups for the hands, whereas mine is based on hands, and doesn't show the game tree, mainly because it's so complicated. I really need to completely rewrite the gui in another language, but really, I just can't be arsed.
Did you ever get out the one with rake? Any plans to pump out some more files? I'd really like to see something with a mono flop.
I did add the rake, and I need to put out an update to the gui so that other people can look at the new solutions. I've rather taken a break from it over Xmas because I added fractional ranges, but I just can't get the exploitability calculation right, which I found rather demoralizing. I've been working on this for exactly six months now, and I'm starting to run out steam I think. However, what I have been doing is systematically finding some solutions for my UTG range v my BTN calling range (it's all about me, me, me now!) on various flops. Once I've finished that, which shouldn't take too long this month, I'll release a new gui and all the files, and let Jules loose on them.
Now, Xmas presents!
I got the book because I enjoyed the film so much. Tbh, if they'd have made the film and called it something else, I don't think the author would have been able to sue. I've never read a book so unlike the film. The book was very good, but the film was infinitely better imo. Not sure I've ever said that before.
Very entertaining. Well worth reading.
TT kept banging on about this one, so I felt I ought to read it. It was as good as I expected, but I have to say, having read The Black Swan as well, I really don't like his writing style. I can't help thinking he's smarter than me, but not as smart as he thinks he is.
Read it, lost money at Zoom PLO10. Watched Skoldlivejc's videos. Lost some more money. I think I've fallen in love with PLO.
I also got the new Pineapple Thief album, which is much better than their last two,
and the new NAO album, on which the jury is still out. I can never get my head around these until I've heard them three or four times,
I also got the new Opeth album, but can't motivate myself to listen to it. A colleague of mine, who's a bit of a closet metal head, went to some ghastly festival last year where Opeth were playing, and said they just played their old songs.
Finally, I've been thinking that, having learnt all this game theory, I really ought to find some way of integrating it into my professional life, so my new year's resolution (I don't believe in new year's resolutions, so it's really just a coincidence) is to try to find a problem that brings together my expertise in differential equations with my superficial knowledge of, and enthusiasm for, game theory. Evolutionary game theory sounds promising, as does the idea of a game that involves partial differential equations. And you're right....I have no idea what I'm talking about, which is always the best place to start.
Happy New Year to my long-suffering readers.
I wonder it the movie starring Scarlett Johansson had anything to do with why it trumped the book? I may have to check it out just to see. For science.
I've been trying to think of other fields to apply these GT algos too. But as of yet, resultless. I'm sure they exist and could be (may already be) used for some pretty exciting stuff. I keep getting stuck on that I think that one needs a very specific problem. I don't know, maybe it could be really useful as a thinking tool for a smaller part of a problem that could somehow be used for the whole of multiple problems. Make sense? Probably not... Possible? no idea... 
spoiler on a rant re taleb
disclaimer: I've been on an email list run by V. Niederhoffer for ~15 years now. I'm well on the VN side of the camp.
Apart from being a a bit of a pretentious prat, Taleb is an unsuccessful money manager. His emprica fund shut down after a few years, coz it was losing money. Bet that wasn't in the blurb. Maybe it didn't spectacularly blow up. However death by a thousand cuts still accomplishes the same thing as a knife through the heart. Same difference.
At least VN has ~4 decades of above market returns and a decent approach to back it up.
Taleb is an opportunist. Or more likely, dumb and lucky enough to be in the right place at the right time.
Black Swan/FbR & Taleb provided the perfect spin for zeitgeist peddlers who wished to blame not so fun financial events on hubris, greed & mis marketed voodoo (instead of the rational outcome of governments stupidly fucking with credit flows). Easy marketing on populist , lackluster content.
But who cares, when you can make 50k per speaking engagement lecturing to some inept money managers. Money managers that are paid by commissions instead of being competent. Who can then go name drop to their buddies about who they spent lunch with. Good work if you can get it.
He's probably taken the 'shit happens' mantra about as far as it can be taken. Personally, I think 'shit happens' holds the equivalent amount of wisdom in a more simple & pleasant form. Plus it doesn't overstep the mark, to something like 'shit always happens'. Even if, sometimes it may feel like it does.
He has another more recent book, called 'antifragile: things that gain from disorder' or something, which is another exercise in self-aggrandizing narcissism. Maybe people who enjoyed his others will like this also.
Pretty similar to his others. Banal & often dangerously false content that relies on his readers getting value by sharing in his elitism. Much like an episode of Jerry Springer or Judge Judy, but without the excitement of a midget love triangle.
Harsh enough?
drops mic.... 
I think it's only fair one reads VN's 'practical speculation' after black swan to get the other side of the story. I'd probably recommend his other book too, but it was so long ago I read it, I can only really remember something about checkers... 
Originally posted by jules97
I wonder it the movie starring Scarlett Johansson had anything to do with why it trumped the book? I may have to check it out just to see. For science.I've been trying to think of other fields to apply these GT algos too. But as of yet, resultless. I'm sure they exist and could be (may already be) used for some pretty exciting stuff. I keep getting stuck on that I think that one needs a very specific problem. I don't know, maybe it could be really useful as a thinking tool for a smaller part of a problem that could somehow be used for the whole of multiple problems. Make sense? Probably not... Possible? no idea...
spoiler on a rant re taleb
Spoiler
disclaimer: I've been on an email list run by V. Niederhoffer for ~15 years now. I'm well on the VN side of the camp.
Apart from being a a bit of a pretentious prat, Taleb is an unsuccessful money manager. His emprica fund shut down after a few years, coz it was losing money. Bet that wasn't in the blurb. Maybe it didn't spectacularly blow up. However death by a thousand cuts still accomplishes the same thing as a knife through the heart. Same difference.
At least VN has ~4 decades of above market returns and a decent approach to back it up.Taleb is an opportunist. Or more likely, dumb and lucky enough to be in the right place at the right time.
Black Swan/FbR & Taleb provided the perfect spin for zeitgeist peddlers who wished to blame not so fun financial events on hubris, greed & mis marketed voodoo (instead of the rational outcome of governments stupidly fucking with credit flows). Easy marketing on populist , lackluster content.
But who cares, when you can make 50k per speaking engagement lecturing to some inept money managers. Money managers that are paid by commissions instead of being competent. Who can then go name drop to their buddies about who they spent lunch with. Good work if you can get it.
He's probably taken the 'shit happens' mantra about as far as it can be taken. Personally, I think 'shit happens' holds the equivalent amount of wisdom in a more simple & pleasant form. Plus it doesn't overstep the mark, to something like 'shit always happens'. Even if, sometimes it may feel like it does.
He has another more recent book, called 'antifragile: things that gain from disorder' or something, which is another exercise in self-aggrandizing narcissism. Maybe people who enjoyed his others will like this also.
Pretty similar to his others. Banal & often dangerously false content that relies on his readers getting value by sharing in his elitism. Much like an episode of Jerry Springer or Judge Judy, but without the excitement of a midget love triangle.Harsh enough?
drops mic....
I think it's only fair one reads VN's 'practical speculation' after black swan to get the other side of the story. I'd probably recommend his other book too, but it was so long ago I read it, I can only really remember something about checkers...
Naked Scarlett Johansson is merely the icing on the cake of a real work of art imo.
Re Taleb rant - well, you are certainly very exercised by him. You have to remember that any opinion I have is from the point of view of a mathematician/scientist who isn't even slightly interested in investing or finance. My business is creating mathematical models of the real world, and as such I'm forever trying to question my assumptions and compare to reality. I don't want to end up like those economists that Taleb laughs at. Taking randomness into account in deterministic models (uncertainty quantification as its called now) is a hot topic here at Nottingham and elsewhere. Personally, I've always tried to steer clear of stochastic models because I know fuck all about them, but poker has given me more of an interest. I also have a PhD student who is developing a stochastic model of an etching process, for which we're having to talk to somebody else because I'm not in the business of bullshitting my students. A lot of my colleagues are interested in problems like nuclear waste disposal in permeable reservoirs for which there's limited data on the permeability. In that context, stock market crash = poisoning people's water - best taken seriously.
On balance, I'm happy for someone to be pushing the philosophical agenda that Taleb's pushing, even if he's not a great investment manager, simply because I'm not interested in finance. I've always had the philosophy that models that involves cells and/or genes are best avoided, which has kept me away from Mathematical Biology, so anything that involves people, who, with some notable exceptions, are less easy to model that either cells or genes, is also beyond my remit.
