Originally posted by EmanuelC16
Originally posted by jbpatzer
Originally posted by EmanuelC16
Originally posted by TwiceT
The yt video hasn't been edited in the German forums ... so if you quote the German post with the link here ... is it still against the forum rules?
The English are a tad more sensitive than the Germans when it comes to critique. They go with the principle that everyone is allowed to speak freely as long as they agree with us. What could go wrong?
As an Englishman, I take that personally, and it's clearly a xenophobic statement, making sweeping, unjustifiable assumptions about both the Germans and the English. Can a moderator please delete this fucking idiot's post?
FYP.
Good point smartarse. Apparently I can't even write English any more. Bloody Romanians.
If personal abuse is not tolerated here, how did this blog get 400 odd pages?!
This black member change is no surprise to me. PS/playtech long ago joined the trend of shunning profitable poker players and focusing on gamblers as their prime market. The overhead for PS must be getting very, very small. No hand reviews, less quality videos, less coachings, no black member. One wonders exactly where the money going? Good news is that there really isn't much more content they can cut, because all the good stuff is pretty much contributed freely by members now anyway.
Congratulations on getting some interest in your code. It's about time. I suppose you probably already know Alex has been getting out some flop solutions. I took up an offer for 100/month to take part in his flop solutions alpha. He has an already done 50-100 flop solutions, will take 10 requested solutions and fast increasing solution library. I'd imagine once it goes public in beta price should increase a bit.
Hopefully he gets some forums up and there's a few willing participants to discuss solutions.
Honestly the biggest detriment to all this is one still has to do the work... You guys should build some kind of 'candy crush' pavlovian reward (or perhaps a punishment would appeal more....) into your solutions or something. That way I might actually get something done. Alex said he's going to get a forum up to discuss his solutions in a week or so. Active discussion of solutions would be nice. But I'm not holding my breath.
Interestingly hesaid he's got solutions (of about 15% v 15% at 1% error) down to about an hour and thinks he may have a big speed boost coming.
Having last night ripped the medial ligament in my right knee, 
and lost the ability to walk for a while, my vegas trip has been postponed and I should have a nice combination of more time mixed with less excuses to grind through a bunch of solutions. Which will most likely result in feeling even worse about myself when I get nothing done. 
Originally posted by jules97
If personal abuse is not tolerated here, how did this blog get 400 odd pages?!
A very good question.
England are rubbish at one day cricket. This World Cup was always going to be a complete disaster.
Congratulations on getting some interest in your code. It's about time. I suppose you probably already know Alex has been getting out some flop solutions. I took up an offer for 100/month to take part in his flop solutions alpha. He has an already done 50-100 flop solutions, will take 10 requested solutions and fast increasing solution library. I'd imagine once it goes public in beta price should increase a bit.
Hopefully he gets some forums up and there's a few willing participants to discuss solutions.
Honestly the biggest detriment to all this is one still has to do the work... You guys should build some kind of 'candy crush' pavlovian reward (or perhaps a punishment would appeal more....) into your solutions or something. That way I might actually get something done. Alex said he's going to get a forum up to discuss his solutions in a week or so. Active discussion of solutions would be nice. But I'm not holding my breath.Interestingly hesaid he's got solutions (of about 15% v 15% at 1% error) down to about an hour and thinks he may have a big speed boost coming.
Having last night ripped the medial ligament in my right knee,
Yes, interpreting the solutions is certainly the hard part.
I don't know how he's managing that in an hour. Hats off to him. Maybe he's multithreading. All the cool coders multithread. I want to multithread too.
Sorry to hear about your knee. GWS
wtf is going on!!!!!
Originally posted by roopopper
Originally posted by Dublimax
Originally posted by GrandeSmokio
just wanted to inform you guys that UPAY4DINNER is read only noweverybody who wants more information visit 2+2 or inform yourself in the thread of the german community
Let the censorship continue...
Originally posted by Dublimax
Originally posted by GrandeSmokio
just wanted to inform you guys that UPAY4DINNER is read only noweverybody who wants more information visit 2+2 or inform yourself in the thread of the german community
Let the censorship continue...
Here is the uncensored thread on 2+2 with a Pokerstrategy rep:
http://forumserver.twoplustwo.com/29/news-views-gossip/pokerstrategy-cancelling-black-level-vip-member-events-1511655/Originally posted by Dublimax
Originally posted by GrandeSmokio
just wanted to inform you guys that UPAY4DINNER is read only noweverybody who wants more information visit 2+2 or inform yourself in the thread of the german community
Let the censorship continue...
Here is the uncensored thread on 2+2 with a Pokerstrategy rep:
http://forumserver.twoplustwo.com/29/news-views-gossip/pokerstrategy-cancelling-black-level-vip-member-events-1511655/Here is the uncensored thread on 2+2 with a Pokerstrategy rep:
http://forumserver.twoplustwo.com/29/news-views-gossip/pokerstrategy-cancelling-black-level-vip-member-events-1511655/thanks for the info will read with interest
these guys claim to have flop solutions down to a few minutes!
http://piosolver.myshopify.com/
Originally posted by jules97
these guys claim to have flop solutions down to a few minutes!
http://piosolver.myshopify.com//blockquote >O bloody hell. Pipped at the post. Don't see how they can do it in a few minutes, but there you go.
Was pretty pissed of about Piotr and his sexy solver last night. I was just about to get a reasonable amount of money from selling my code. I'm over it now. As I just said to Manu, I basically bubbled a freeroll. I didn't write it for the $$, but for the last three weeks I have been coding like a demon trying to make it saleable. I've learnt lots of stuff that I'll be able to use in the code that I often have to write for a living, so that's all good. I have a friend who's going to get Piosolver and give me info about runtimes and memory usage in return for checking solutions against my code. If they're able to do big problems from the flop in minutes, they have some innovation that I don't, and interestingly that the University of Alberta don't either, as far as I know. If I find out, or work out, what it is, I'll be on it like a rat up a drainpipe. I also think that this makes the solution of a full HUNLHE game with restricted bet sizes possible for someone with the sort of computing resources that were used to solve HUFLHE, assuming it's kosher and not a scam. I suspect the former is true, but I'm buggered if I know how they're doing it. They do mention gradient descent methods on their website, but it seems implausible for such big problems. Anyway, if Piosolver turns out to be as good as it says it is, I'll release my code for free in all its glory. That will save you $250 at the expense of having to wait for longer for solutions.
If they're able to do big problems from the flop in minutes, they have some innovation that I don't, and interestingly that the University of Alberta don't either, as far as I know. If I find out, or work out, what it is, I'll be on it like a rat up a drainpipe. I also think that this makes the solution of a full HUNLHE game with restricted bet sizes possible for someone with the sort of computing resources that were used to solve HUFLHE, assuming it's kosher and not a scam. I suspect the former is true
All of this is correct. We are willing to make it as easy as possible for other tools to verify our results. Everything is loggable/possible to print to file with command line interface. We are happy to provide more tools for verifiying/testing of our results for people who are serious about it. Just contact us at piosolver@piosolver.com
Regards,
Piotrek
Originally posted by Punter11235
If they're able to do big problems from the flop in minutes, they have some innovation that I don't, and interestingly that the University of Alberta don't either, as far as I know. If I find out, or work out, what it is, I'll be on it like a rat up a drainpipe. I also think that this makes the solution of a full HUNLHE game with restricted bet sizes possible for someone with the sort of computing resources that were used to solve HUFLHE, assuming it's kosher and not a scam. I suspect the former is true
All of this is correct. We are willing to make it as easy as possible for other tools to verify our results. Everything is loggable/possible to print to file with command line interface. We are happy to provide more tools for verifiying/testing of our results for people who are serious about it. Just contact us at piosolver@piosolver.com
Regards,
Piotrek
I've been thinking about this, and I reckon part of the explanation is that I'm doing something stupid somewhere with my algorithm. I just don't believe that you, however clever you are, and I'm sure you're really very clever indeed, can have found an innovation that makes your solver run 3-4 orders of magnitude faster than mine. By far the more likely explanation is that I am an idiot, and, proceeding on that basis, I think I've found a moron error in my code. I have the right answer, but I think I'm doing at least two orders of magnitude more work than I need to, which would explain a lot (story of my life). Anyway, I will definitely take you up on your offer, but I will try to sort out my solver first.
Shame about just getting beaten on the finish line. It seems his solver is gaining some popularity. Must be the power of 2plus2. Found out yet if you did something stupid?
You should release something anyway. Undercut the bastards, put it out for 20 or 50 bucks or something. And make the solution viewer open source so other people can fix that up and your laughing.
I just imagine all these guys new to these solutions, going 'fuck these mixed strategies'.
Whether or not it's 5 minutes or a day is mostly insignificant to me. At about one a day, I can keep up with time spent analyzing them. One every 5 minutes is terrifying. GTORB has the best interface and I think that's more important. He's got a nice library of growing solutions too now. And what's better is that other people do all the work of submitting and building them. 
But I think I'll take a break for a few months from going over these. Getting a decent feel for the solutions, plus my head hurts and I've probably spent more time and effort on these for now than is efficient.
Originally posted by jules97
Shame about just getting beaten on the finish line. It seems his solver is gaining some popularity. Must be the power of 2plus2. Found out yet if you did something stupid?
You should release something anyway. Undercut the bastards, put it out for 20 or 50 bucks or something. And make the solution viewer open source so other people can fix that up and your laughing.I just imagine all these guys new to these solutions, going 'fuck these mixed strategies'.
Whether or not it's 5 minutes or a day is mostly insignificant to me. At about one a day, I can keep up with time spent analyzing them. One every 5 minutes is terrifying. GTORB has the best interface and I think that's more important. He's got a nice library of growing solutions too now. And what's better is that other people do all the work of submitting and building them.
But I think I'll take a break for a few months from going over these. Getting a decent feel for the solutions, plus my head hurts and I've probably spent more time and effort on these for now than is efficient.
It turns out I was doing something mildly silly, which I fixed and got a speedup of about a factor of 3-5, but his claim is that he really is doing something genuinely at least 1000x faster than CFRM. He's dropped me some hints about what he's doing, and it sounds like something that I've used on toy games, but which I just assumed wouldn't work on such a big game. I'm working on giving it a go now. Serves me right for looking for the state of the art in the literature instead of just trying stuff I've used before. The ironic thing is that all the best things I've done in my career have been when I've just decided to ignore all the literature and plough ahead anyway. 
I think that within the next year it'll be obvious to everybody how to solve these problems and people will realize that (i) having a great GUI is a key selling point of any software (ii) providing some understanding of the solutions is hard...much harder than just obtaining them. I think on both those points that GTORangeBuilder will do very well, for the reasons that you suggested.
Originally posted by jbpatzer
I think that within the next year it'll be obvious to everybody how to solve these problems and people will realize that (i) having a great GUI is a key selling point of any software (ii) providing some understanding of the solutions is hard...much harder than just obtaining them. I think on both those points that GTORangeBuilder will do very well, for the reasons that you suggested.
Ah so, an open source viewer isn't the model then. In a years time the solution solver won't be the intellectual property, the viewer will. Nice.
Yeah I had a chat to Alex about the flop solution packs he's started selling. Where if you haven't seen them, they're 10 or so solution and a video of him providing some analysis. He said interest in the solution packs compared to actual solution is someting like 10 to 1. Which makes sense. It's a lot mroe fun to watch a video of someone smart telling me what's what than grinding out each individual combo.
Will be interesting times for NLH, but I think most people underestimate the difficulty of applying the principles of GTO.
I actually find the solutions fairly straightforward, perhaps more so for FL than for NL. Especially now with just 1 fixed betsize most of the hands have to follow a somewhat logical pattern.
If anything I see this hurting 90-95% of the guys. I think the current way of displaying things is not very efficient. I have some ideas/stuff I and few others are working on but yea I kind of feel like Jules lately with this shit just hurting my head.
Originally posted by jbpatzer
Originally posted by jules97
England are rubbish at one day cricket. This World Cup was always going to be a complete disaster.
Hmm. Just popped in to reminisce about ol' times in the blogging dept and instantly chastised myself for not making the effort to come here sooner and dig the knife in. I must commend you on your implied die hard optimism about forms of cricket outside the ODI format. I give that... when are we touring again?... right, about 2 months.
Take care old bean.
Originally posted by Waiboy
Originally posted by jbpatzer
Originally posted by jules97
England are rubbish at one day cricket. This World Cup was always going to be a complete disaster.
Hmm. Just popped in to reminisce about ol' times in the blogging dept and instantly chastised myself for not making the effort to come here sooner and dig the knife in. I must commend you on your implied die hard optimism about forms of cricket outside the ODI format. I give that... when are we touring again?... right, about 2 months.
Take care old bean.
Well, you know English cricket goes up and down - one year we win the 20-20 World Cup, the next we're kicked out of the 50 overs World Cup, one year we're number one in the Test rankings, the next we're getting pasted by Australia - which keeps it interesting. You can keep the relentless 'almost the best, but never quite actually getting there' flavour of NZ cricket. I'll stick with England for entertainment value. Remember I have the experience of supporting Spurs for > 40 years.
As you can tell, I've not been playing much poker. I've been talking to Piosolver Piotr a lot and I've been giving him some help whilst he gradually tells me about his solver. I remember when I first got my solver working, my aim was to be able to do the 10 example hands in Janda's book. The last time I wrote anything here, each of them took several hours to calculate, but I could do it. I now have a code that can do all 10 in less than 10 hours. With Piosolver, it would probably take less than an hour (I haven't tried, but I could because Piotr was nice enough to give me a free copy of Piosolver
). I think I now know how to do it as fast as that and can aim for faster (there's some friendly rivalry here). So, I'm still working on my code, but it's never going to be a commercial product. It'll just be for my own use.
My other main news is that mrsjb went and bought a puppy (and note that I'm now blogging in C, not MATLAB)
for (int i = 0; i<=10000000; ++i)
{
Mrsjb: Can we get a dog?
Me: No, I hate dogs.
Daughter: Can we have a puppy?
Me: No we can't.
}
Mrsjb: Can we get a dog?
Me: O ffs, I suppose so.
.
.
10 seconds later
.
.
Mrsjb: We have a puppy!
Me:
She's alright...for a dog. The fact that she's completely mental makes her a good fit for my family.
Originally posted by jbpatzer
My other main news is that mrsjb went and bought a puppy (and note that I'm now blogging in C, not MATLAB)
#! /usr/bin/env python2
""" to prevent acquisition of cat
"""
while True:
If asked == 'May we get a kitten':
answer = 'No!'
If visiting_grandma and kitten_cuteness == 10 and kitten_loves_Vorpal:
break
cat.name = 'Punkin'
cat.favourite_perch = Vorpal.shoulder
cat.home = Vorpal.ManorEnjoy your puppy before it grows up!
Peace,
VS