"True M" vs. Harrington's M: Critical Flaws in Harrington's M Theory

[stevepa]

[ QUOTE ]
Steve,

I hesitate to try because you are, if not smarter than me, definitely a better poker player. I don't always agree 100% with OP, although I do on most things. He may also stop by and tell you my explanation is wrong. Now that I've got the disclaimers out of the way I'll give it a shot. [img]/images/graemlins/smile.gif[/img]

[/ QUOTE ]

I appreciate the effort [img]/images/graemlins/smile.gif[/img]

[ QUOTE ]
What is a +cEV play? Anything that on average will win more chips than we lose, right? OP doesn't suggest we make -cEV plays since, by definition, continually doing that guarantees you'll lose.

[/ QUOTE ]

Well yes and no...there are some spots which some people will argue are -cEV but +$EV. I'm still on the fence about whether I think those exist but was unsure if that's why the author was arguing or not.

[ QUOTE ]
What he suggests is that you look for your +cEV plays that aren't always based purely on your cards. Is a blind steal +cEV? How about calling a pre-flop raise (with positon) from a player you've identified as weak tight with the intention of taking away the pot on the turn if he shows any weakness? I'm sure you've done both of these at some point.

[/ QUOTE ]

Everybody does these things and I'm sure if you asked Harrington he'd agree as well. But whether those are profitable plays in a given hand is based on: a) your cards (which always have at least a minor importance) b) effective stacks c) your opponents tendencies d) respective images, players to act and their stacks/tendencies, flow of the game, etc. etc. but NOT how soon the blinds go up. (I'll come back to this)

[ QUOTE ]
If Harrington really believes you should always take all +cEV edges then why would he suggest when you get short stacked that you have to loosen up your starting requirements. Does the hand that isn't good enough with a bigger stack suddenly become profitable due to you having less chips?

[/ QUOTE ]

Yes it does. Hand values change substantially as stack sizes change. A9o is complete trash when you have 50bb's but a pretty solid hand when you have 5. I do think Harrington tends to err on the tight side, but if I remember correctly his approach is to take anything he perceives to be +EV.

[ QUOTE ]
We make adjustments due to a quicker structure because we are (or should consider ourself to be) desperate quicker than using Harrington's zone system would indicate.

[/ QUOTE ]

Does this mean you're pushing hands earlier? Playing more hands? If so, why aren't you playing these hands in the slower structure? Are you purposely passing up +cEV plays in the slow structure to "wait for a better spot"? If so, why? I feel like what you're actually arguing is that Harrington plays too tight. Maybe that's true but how tight you should play is based so much on table dynamics that it's not really a fair criticism.

[ QUOTE ]
[ QUOTE ]
P.S. I strongly disagree with the idea that tournament structure has a substantial impact on strategy in all but the most extreme cases.

[/ QUOTE ]

Frankly I don't see how anyone can believe this. Possibly if you read Snyder's book instead of basing your opinion on a quick browse in the bookstore you'd feel differently. Then again, maybe not.

Al

[/ QUOTE ]

I can't really have a firm opinion on Snyder's book because like you said, I haven't really read it. What I did read had some clearly wrong ideas (e.g. calling any two on the button to a raise) but maybe he addresses those later. Anyways, that's not really the point of this thread.

Here is a very quick explanation of why I don't believe structure matters in tournaments. Like I said before, whether or not playing a hand is profitable depends on a lot of factors, including current blinds, chips stacks, images/abilities of others and yourself, etc. But the fact that the blinds are going to go up in 5 minutes instead of 10 certainly doesn't change the cEV of a given play. So the only arguments in favour of structure mattering is that in fast structures you have to take -EV spots or in slow structures you should pass up +EV spots. I think there are spots where you should take -cEV spots if the blinds are about to go up if you'll lose your FE upon the blind increase. But these spots happen in all tournaments, regardless of structure; they just happen more often in fast structured ones. Passing +EV spots has been discussed many times, I think there are very few times it's correct. But regardless, your goal in every hand is still to maximize EV (where cEV is generally a very good approximation of $EV). How soon the blinds go up very rarely if ever impacts EV and thus structure of tournaments is irrelevant to proper strategies.

Steve

[betgo]

But isn't the main aspect of M your stack size relative to the pot, rather than the number of rounds till you are blinded out. I just don't see the point to the article.

[BigAlK]

[ QUOTE ]
But isn't the main aspect of M your stack size relative to the pot, rather than the number of rounds till you are blinded out. I just don't see the point to the article.

[/ QUOTE ]

Sure M is computed by figuring the ratio of your stack size to the pre-flop blinds and antes. But Harrington specifically says the reason why this figure is important is that it indicates how long until you'll be blinded out (I think Snyder has the exact quote and page number in his article). As I read it this "how long have I got" aspect of M and therefore how aggressive do I need to play to remain competitive is one of two uses Harrington uses M for. The other use Harrington outlines is in deciding which hands are playable (ie, small pairs go down in value in the yellow/orange zone because you can't play them for set value).

[Dr1Gonzo]

[ QUOTE ]

(ie, small pairs go down in value in the yellow/orange zone because you can't play them for set value).


[/ QUOTE ]
allegedly! You can of course bring implied odds into this and say you can always call with a small pair assuming it is for only an eigth of your stack.
I suppose that's besides the point though.

If no alternative has been offered to calculating M or an efficient way of doing this at the table then it's effectively business as usual. Some people use M to tell them when to get aggressive, others just play aggressively anyway.

[betgo]

[ QUOTE ]
[ QUOTE ]
But isn't the main aspect of M your stack size relative to the pot, rather than the number of rounds till you are blinded out. I just don't see the point to the article.

[/ QUOTE ]

Sure M is computed by figuring the ratio of your stack size to the pre-flop blinds and antes. But Harrington specifically says the reason why this figure is important is that it indicates how long until you'll be blinded out (I think Snyder has the exact quote and page number in his article). As I read it this "how long have I got" aspect of M and therefore how aggressive do I need to play to remain competitive is one of two uses Harrington uses M for. The other use Harrington outlines is in deciding which hands are playable (ie, small pairs go down in value in the yellow/orange zone because you can't play them for set value).

[/ QUOTE ]
I agree that M overestimates how long you have before being blinded out, but I don't think that is the main point of M. Who folds all their hands short stacked anyway. Certainly that is not how Harrington advises you to play.

Snyder makes a point, but it is a really trivial irrelevant point. This guy shows very little understanding of strategy in his writings, but he makes it out like this is some great refutation of Harrington or Sklansky.

[Sherman]

[ QUOTE ]
What are these adjustments that we make when the structure is worse?

[/ QUOTE ]

Imagine two tournaments. One with 5 minute blind levels and one with 30 minute blind levels. If you know that you are only going to get 1 hand per minute (let's be generous) obviously the 30 minute structure is better for a good player. No one will argue that.

So should we play differently in the 5 minute structure than in the 30 minute structure? I think the answer is, "For the most part no, but for some parts yes."

For example, if your M is 10, you should play more aggressively in the 5 minute structure because you know that you won't get as many opportunities before your stack is too short to do anything. Let's assume the blinds double at each level. In the 5 minute structure you get to see 5 hands per level. That means if you fold all 5 this level, your M is now 5 (even if you didn't pay the blinds, if you did it is 3.5). And in 5 more hands, your M will be 2.5.

So, a tough situation might come up. Let's say you get dealt 77 UTG. What do you do? In the five minute structure, it is more likely that a push is correct because mathematically, it is less likely that you will get a better opportunity. In a 30 minute structure with an M of 10, you can easily fold 77 UTG and not worry because you have 29 more hands at this level and 30 more at the next. Thus, the likelihood that you will find a more profitable situation is greater.

This example probably isn't mathematically perfect, but I think it demonstrates what Snyder is talking about. Which is that sometimes structure matters, because a more profitable situation may not be very likely to come up with limited time.

Now, his point is that given that structure matters, one should take advantage of profitable situations when they arise (no conflicts here). As opposed to my example above, he suggests that people use things like position and their chip stack to create profitable situations and to take advantage of them before it is too late. The details of which are covered in his book.

I own and have read his book. It is decent. I have tried employing his strategies. I didn't find them very useful. I found that they put me in situations where I wasn't comfortable playing and I often made mistakes. Which seems to go against his whole argument of finding profitable situations. Anyhow, I hope some of this post makes sense.

[seke2]

To me, OP's article just comes over as being a nit about reading Harrington's definition of M overly literally and assuming that poker players are just robots incapable of understanding anything deeper about it.

Like everyone here has said for a long time, M is just general guide, it's not the be-all-end-all of poker knowledge. M helps you quickly understand where you stand in relation to the blinds/antes.

I leaned moderately heavily on M when I was first learning to play well. Soon, I evolved to the point where the things that M told me became implicit, and now I mostly just think in terms of BB's because it's an easier way for me to figure where I am, and I can just automatically take into account the ante structure.

So to me, Snyder just comes off like one of those guys who sees a post on a message board where there are a handful of minor grammatical mistakes and refutes the other person's post on the basis of a few typos and misplaced commas. I didn't get anything out of reading, really. We all know that "True M" is a relative measure which is impacted by many things, and taking the time to calculate and understand "True M" is ridiculous. I'm willing to bet most of the better players on this forum tend to think inherently in terms of stack size in BB's as opposed to even thinking of stack size in terms of M, anyway.

[betgo]

There are some adjustments you need to make with a very fast structure. Harrington made an approximation that M was the number of rounds until you are blinded out. This is an accurate approximation for the tournaments which Harrington plays in, with 90+ minute rounds, but it is not accurate for low buyin live and online tournaments with very fast structures.

In some tournaments with very fast structures, you may need to gamble to maintain a stack large enough that you will have folding equity when the blinds increase. Most of the books assume major tournaments with slow structures, and do not discuss strategy for turbo tournaments.

[DP388]

[ QUOTE ]
So the only arguments in favour of structure mattering is that in fast structures you have to take -EV spots or in slow structures you should pass up +EV spots. I think there are spots where you should take -cEV spots if the blinds are about to go up if you'll lose your FE upon the blind increase.

[/ QUOTE ]

I'm still on the fence in regards to whether tournament structure means something. But I do know that, for example, if the blinds are about to change from 100/200 +25 to 200/400 +25 and I have 3000 in chips and I'm faced with a decision to re-steal I am more apt to push on the resteal knowing that I need chips to stay ahead of the next round.

Maybe this is incorrect thinking and it shouldn't matter because the +EV of the re-steal is there regardless. But if I'm on the fence, this might be the thing that tips the scale.

Again, this might be incorrect thinking. Thoughts?

[BigAlK]

[ QUOTE ]
Everybody does these things and I'm sure if you asked Harrington he'd agree as well. But whether those are profitable plays in a given hand is based on: a) your cards (which always have at least a minor importance) b) effective stacks c) your opponents tendencies d) respective images, players to act and their stacks/tendencies, flow of the game, etc. etc. but NOT how soon the blinds go up. (I'll come back to this)

[/ QUOTE ]

[ QUOTE ]
Does this mean you're pushing hands earlier? Playing more hands? If so, why aren't you playing these hands in the slower structure? Are you purposely passing up +cEV plays in the slow structure to "wait for a better spot"? If so, why? I feel like what you're actually arguing is that Harrington plays too tight. Maybe that's true but how tight you should play is based so much on table dynamics that it's not really a fair criticism.

[/ QUOTE ]

Snyder is definitely saying that Harrington is too tight. I agree (and believe Snyder does too) with the things I quote above. His theories are largely based on playing against a large number of people playing the strategies that Harrington proposes. This apparently worked well for him in the low buy-in B&M tournaments he plays in Vegas. The opponents and table dynamics he found in these games allowed his techniques to work because he felt those playing a Harrington strategy could be pushed off a hand easier or more often. I've found that what works best for me is usually somewhere between Harrington and Snyder for the games I play depending on all the factors you list. For example the calling with any two on the button suggestion is something I rarely use. When I do it's not normally any two and is dependent on who else is in the pot.

[ QUOTE ]
But the fact that the blinds are going to go up in 5 minutes instead of 10 certainly doesn't change the cEV of a given play. So the only arguments in favour of structure mattering is that in fast structures you have to take -EV spots or in slow structures you should pass up +EV spots.

[/ QUOTE ]

I'll throw out one potential example. Let's say blinds are 200/400 and your stack is 7200 (18BBs or an M of 12). Assume you're in utg+1 and utg folds. You have 77.

If I interpret Harrington correctly he would advise mucking here. In the section on yellow zone play he discusses playing 44 from utg+1 and concludes that you can't play it in this situation. Although 77 is slightly stronger I believe the logic he outlines would still apply.

On the other hand Snyder's advice in this situation would be to go all-in. He says, "when you're this short on chips you must take risks because the risk of tournament death is greater if you don't play than if you do." In other words with blinds rising rapidly the odds of you getting a better hand in a better situation to pick up a few chips may not come.

Obviously Harrington's approach is neutral cEV. I guess whether Snyder's is + or - cEV depends on how you view your opponents calling ranges and the probability of one of those yet to act having one of these hands. I see Shermn27 just made a post with what at first glance looks like about the same example so I won't attempt to do the math here.