>A few years ago, a young pro named Phil Galfond published a crucial refinement to Mr. Sklansky's point. He showed that the right way to analyze a poker decision is to consider your opponent's "range"—that is, the full set of different hands that he could plausibly have, given all the actions that he has thus far taken.
I find it impossible to believe that concept of a "range" was (a) invented a few years ago and (b) invented by Phil Galfond. This article is disappointing.
Edit: In fairness, the article's sentence is literally correct. But I cringe at its implication.
I have a theorem which I refer to as the Fuzzwah theory, it goes like this:
"The more you know about a topic, the more it will annoy you when it is covered by the main stream media."
Once you realize this with regards to topics you're knowledgeable about you start taking coverage of ALL topics with a grain of salt, knowing that someone who knows all about it is probably cringing and / or yelling at how poorly it is being reported.
PS: I was obsessed with online poker a few years back and was an avid 2plus2 forum reader. The 1st person which I recall giving excellent advice about putting people on a range was Cole "CTS" South. He even sold an ebook called Let There Be Range for (initially, from memory) something like $5,000 a pop (it still sells for $1400 on amazon http://www.amazon.com/dp/0982402252 )
Mainstream media? I occasionally cringe at some of the discussion here on subjects I know something about, seeing how badly wrong smart people can get it. (Thankfully most of the discussion on HN is on subjects I know little about, or are that are way over my head! :-) )
What is your point? No one can be smart at everything. Just because one occasionally writes words that helps computers do things doesn't mean one's world view is not silly when it comes to other things.
Maybe we should organize a hacker news charity poker tournament. Nothing big, ins are $100, winner gets $200, 2nd place gets money back, and everything left goes to Watsi.
Yes, this article is shockingly reductionist -- even downright incorrect -- to anyone who knows something about poker. Run back through the 2+2 archives, people were analyzing plays based on hand distributions since the beginning.
Sklansky's seminal text The Theory of Poker revolves around discussing distributions of hands -- how to guess at an opponent's distribution, how to mix up your own distribution to be harder to read, how to determine optimal (in the game theoretical sense) distributions. TOP was published in 1987. Galfond was 2 years old at that time.
Galfond simply took a concept that already existed and popularized it. He was successful at popularizing because of his idol status in online poker circles, circles in which Sklansky is something of a pariah.
An entire industry has popped up around taking already-existing technical concepts in poker strategy, dumbing them down, and relabelling them. See the rise of video coaching (cardrunners, deucescracked, et al) and more recently poker coaching (wherein people pay 3-figure hourly fees to have pros analyze their play) for examples.
None of these strategy services offers anything new because the math and theory behind poker has not changed. The structure of the game is the same now as it was in the 80s, and the 50s, and 20s, and so on.
The vast majority of modern strategy content is a restatement or application of the content from two canonical texts: the aforementioned Theory of Poker, and Chen's Mathematics of Poker. Even most coaches who charge the three-figure fees admit this. It's just that most people don't want to put in the gritty analysis work, hence the coaches stay in business.
Sklansky is a pariah because of every book he's written other than Theory of Poker. Many of them contain advice ranging from mediocre to atrocious. Also he's the kind of tool who says things like "Ken Jennings must be an idiot because he believes in God."
Excepting some of the timeline issues though, the article is reasonable. Computer analysis radically changed how poker is played, largely by making it more aggressive. Read some books from guys like T.J. Cloutier written before everyone had Pokerstove. They brag about laying down QQ in spots where anyone with a Monte Carlo sim would know it was idiotic to do so.
Most people (even among top professional poker players) can't understand all of Chen's book. Most good poker books go much lower level than ToP, which is useful too, as evidenced by the fact that every time Sklansky himself tries to he says something dumb. If the man who wrote ToP can't apply its concepts correctly, then there's value in other people trying to do so.
>Sklansky is a pariah because of every book he's written other than Theory of Poker
I was thinking specifically of his personal life and how it bled into the strategy forums. For example the whole Brandi debacle, and before that the Sue revelations (remember the pictures with the little girl and the birds?). Many people who read that stuff were so creeped out by it that they discounted Sklansky's strategy talents entirely. So forum people -- who were a sizeable portion, probably a majority, of strategy-conscious online players at the time -- became unlikely to find out that Sklansky wrote at length about 'ranges', creating a gap for someone like Galfond to fill.
I believe that Phil Galfond was one of the first "new-generation" players to explain the strategy with modern media (youtube, podcasts) in a way that it could be taken and applied in your own game without too much hassle.
Harrington had explained the concept of evaluating tournament situations based on a range (although he didn't necessarily call it a range) in his Harrington on Hold 'Em books.
Galfond's gbucks article, maybe not the first mention of it, brought the concept of equity against a range into the mainstream.
Yeah, this is akin to saying that Al Gore invented the internet. A simple trawling of two plus two forums will reveal range discussions from years back.
On a shameless self promotional note: If there's any online casual players out there and they want to know their opponents range, check out a lil web app I built: http://www.pokerstoker.com. Keeps track of opponents pre-flop play, which is essentially their range.
I actually want to add a little feature that displays how your (previous) hand stacks up against various ranges across the streets. I'll add it when the poker bug bites me again. :-)
It's certainly older than a few years. I understood it over a decade ago, and I certainly didn't invent it.
To be technical though, the article doesn't claim that Phil Galfond invented it either. Just that he "published a refinement". He probably posted something on 2+2 that was the earliest example the WSJ's writer found, though he must not have looked very hard if so.
Also computer analysis became a factor before the boom (Chris Ferguson won the WSOP main event largely due to it) but the article is pretty accurate about its impact. It's not too much of an exaggeration to say that nobody was good at poker before the year 2000, or maybe even a bit later. The people who won a lot of tournaments back then would fail miserably if they played the same way today. Some of them do in fact.
I came here just to make this comment. I'm quite sure this concept appears in most of Sklansky's books, probably right back to the Theory of Poker (c1987). (I haven't verified that, but I know his more recent books like Small Stakes Hold'em, still more than a few years ago, covered ranges.)
Confirmed. TOP is one big discussion of ranges. Sklansky didn't call them ranges, but the concept is the same: your opponent holds a distribution of possible hands, and you're making decisions based on what you think the makeup of that distribution is (and also based on what you think your opponent thinks your distribution is).
Also, in 2006 Chen and Ankenman published The Mathematics of Poker which quite explicitly deals with the analysis of hand distributions (e.g. chapter 8: Playing Accurately Part II: Hand vs. Distribution, chapter 9: Adaptive Play: Distribution vs. Distribution).
So even if we pretend Sklansky's TOP doesn't deal with ranges, Chen clearly does, and came before Galfond's 2007 Bluff article [1].
I started playing in 2003 and posted on twoplustwo since 2004. I can certainly confirm that the concept of range was being discussed in that period and I'm actually fairly confident that the Theory of Poker (Sklansky) itself discusses range. The article just seems blatantly wrong in that sense.
There's another element in there. The one about relative skill. Let's say that you're a much better player than the other person, and that you would win 90% of the time heads up (one on one, which we are going to assume at this point for simplicity). If this were the case, you would need to have a greater than 90% chance to win the hand all in for it to be a correct move.
This is also easy to understand if you, a novice, are heads up against one of the pros. It's a much better move for you to go all in on any decent hand and just try to get lucky to win, since if you try to play their game, you're going to lose.
Many poker players who study theory see this as incorrect. Ill try explain the other side.
I will avoid the fact that a 90% winrate in a format like you describe (heads up sit and go) is not even remotely possible vs anyone who knows the basic rules. In heads up sit and goes 2-5% roi is generally considered great at any non micro level. Poker (especially online) is more of a volume game. However, roi is very flawed which I will get to. Generally the most profitable players are ones who crank out a small winrate over large amounts of volume.
Here is the problem though, we really are not concerned with winning a single game in a format like this. Our top concern is making money. In a winner take all heads up tournament chipEV == $EV so we can just call it EV in this format and not worry about ICM. If we are playing for a living we want to make the most amount of money, so each session we want to maximize our EV (win the most amount of chips). Most internet players worry about their winrate in BB/100 or EV adjusted $/hr.
When we get into a situation which we are a 1% favorite for full stacks vs a player we have a large edge on, we should still take it, its money in our pocket long term. If we lose we pray the opponent accepts a rematch. This is especially true for heads up games which usually only last a few minutes and its possible to play 100-200 games per day. A 3% winrate at $100 stakes with this volume becomes a significant amount of money even before factoring in site bonuses (likely a bit over $1/game) etc.
Not related to your comment, but doesn't deserve another post is that poker really is a great game and its unfortunate how its perceived by so many. Ive been playing for 3 years professionally and coached on a popular training site, yet after transitioning to programming (With a cs degree) I get laughed at by recruiters and seen as a degenerate.
He was greatly mistaken in his 90% estimate, but the players with the best winrates when I quit (2011) were still winning ~60% of games.
My ROI was ~14% at 100s when I played professionally, totaling 1000s of games. I agree that at 1ks+ the best players were making ~5% ROI, but there were much better winrates up thru the 500s, which I don't consider "micros."
Also, ROI is generally less at micros, because rake is a bigger issue.
why not? It demonstrates the ability to hold a huge amount of information in your head at once, make rapid decisions in a high pressure environment, and do what are essentially Bayesian probability calculations in your head. I may be biased as someone that played poker to put myself through grad school and had a nice hourly rate by the end of it, but I think that it requires a high intelligence, the ability to emotionally disconnect from decisions, and the ability to retain and act on such large amounts of information that it's a definite positive. (It's also really boring and psychologically brutal at times and I wouldn't recommend it to many folks.)
Are you still a winning player? Were you playing before Black Friday? I heard it was easier. I studied the game a lot these past few months and I barely break even. I'm only winning in live poker.
Not sure...there is a dearth of games now for US players and I was so disgusted with my money getting wrapped up in the Full Tilt BS that I went away from poker for about a year. I am certainly still a winning player at live poker, but the hourly rate can be a bit dismal and I don't enjoy it enough to do it everyday, or even once a week.
I would say that by 2008 the average live game was strong enough that I definitely had to practice disciplined game/seat selection, couldn't just sit down anywhere and expect my edge was higher than the rake.
I agree HU strategies advanced exponentially once stoxpoker and some other video sites came on and people started being heads-up specialists.
I'm hoping the US players can come back online soon. I'm not saying US players are bad but it can't be harder than playing Russians hyper-LAGs all day.
As the field narrows to just the devoted players, it will get harder.
Also, as the article suggests, poker strategies are advancing quickly. I started in 2004, and by 2012 I was still ahead of the curve, but the game had clearly gone light-years forward in terms of general strategy--in particular I'm speaking to HUSNGs, my game of choice.
I agree. I also played professionally for 3 years, and I found that when applying for devloper positions, my poker experience has been lauded as related.
Yes, but then the pros determine that you have a very wide range for going all in. If needed, they will pick a spot where they have a very strong hand to beat what you have, at best an average hand. Additionally, if you're a novice going all in a lot, the whole table will eventually recognize this and all of them will adapt their playing to take your chips. Certainly, you'll get lucky perhaps on a few all ins and win against a pro. And you might actually be right, that this playing style is better than just trying your best as a novice, but it's analogous on a true/false test to just filling in true for every answer. That's only recommended if you know you'll do poorly on the test even if you try. And it's also analogous that playing a pro would be like trying to answer true every time of a professor known to have false be every answer on some tests--i.e., it's a really dumb poker strategy, especially versus pros who are expecting your strategy and thinking one level deeper.
Also, the real reason you're wrong is that if you're heads up against a pro and you go all in, the pro can fold and live to see the next hand.
When I play ring game (non-tournament) poker, I'm attempting to maximize my hourly rate, not minimize my variance. Therefore I am pushing every edge I have, even when small.
>Bluffing still matters, but the best players now depend on math theory
This is a silly byline. If anything, the best and most mathematical players bluff more than the older generation of typically more conservative players. Bluffing and math are NOT opposed principles.
Bluffing is one facet (often romantasised and therefore sometimes over emphasised) of a good mathematical strategy. Like you say they are not mutually exclusive of one a other.
Indeed. There are many situations in which it's mathematically correct to bluff in hold 'em and omaha. Or at least semi-bluff.
I remember playing at the toughest game in my area about ten years ago. Two of the best players (both professionals) were heads up. It was kind of a dull hand after the flop - bet, call; bet, call; check, bet, fold.
As the cards were being collected the winner said "Let's try something different next time. How about the guy with the hand bets and the guy who's on the draw calls."
Reddit has its place, but the level of poker discussion there will never approach that of twoplustwo which is a resource any aspiring player should use
I find it impossible to believe that concept of a "range" was (a) invented a few years ago and (b) invented by Phil Galfond. This article is disappointing.
Edit: In fairness, the article's sentence is literally correct. But I cringe at its implication.