A No-Limit Hold Em Experiment

[Jordan Olsommer]

So I bought Ed Miller's new book, _Getting Started in Hold 'Em_ the
other day, and while I think the entire book is fantastic, one section
in particular intrigued me. On pages 122-135, Miller outlines a system
for making a profit at No-Limit Holdem cash games with a small stack,
by playing "ultra-tight".

It intrigued me so much that I have decided to run a miniature
experiment to see if his small stack system (or as I like to call it,
the "Miller All In Method", or "MAIM" [img]/images/graemlins/smile.gif[/img]) is profitable and if so to
what degree. The results of the experiment will be posted here on a
regular basis,and then a conclusion after a few weeks or once the system
has demonstrated either a ridiculous level of profitability or my
starting bankroll drops to $0 (whichever comes first).

The reason I believe this will be valid as an experiment is because
Miller's system is almost completely algorithmic; there are very few
"judgment calls" outlined in the text, and there are none that are
absolutely required for execution of the MAIM system.

So to find out if the MAIM system works in the Real World (of Internet
Poker), I've set up an account on a poker site with a $300 starting
bankroll. Given Sklansky & Malmuth's advice on bankroll management, I
will only play NLHE games with big bets of $1 or less, thus having the
300BB bankroll they recommend.

Miller states that a small stack is 20-25xBb, so I will buy-in for
20xBb ($5.00 in the 0.25-0.25 games in which I am starting out). **To
avoid any compromising of the experiment, if my stack increases to
30-35xBb or more [Miller's definition of a "Medium stack"], I will
leave the table [can't sit out and then rebuy in for 20xBb again,
unfortunately] - I am doing this to completely isolate the MAIM system
to see if and how well it works; since the MAIM system only outlines
short-stack play, any hands I play with a medium stack cannot
incorporate the system, only my own personal strategic judgments,
which would obviously fudge the results quite a bit (for better or for
worse, I hesitate to guess [img]/images/graemlins/smile.gif[/img]).

Here is the general algorithm for the MAIM system which I will follow
to the letter during the experiment:

#Starting Hand Requirements#

Opening for a raise:

Early Position (1-3): AA-TT, AK
Middle Position (4-6): AA-99, AK, AQ
Late Position: (7,D,Sb,Bb): AA-77, AK-AT, KQ

if raise in front: re-raise all-in with AA-TT, AK
if raise and re-raise: AA, KK
if you get reraised:
reraise all-in when the ratio of money left : your original bet is:
5-to-1+: AA-QQ, AK
4-to-1: AA-JJ, AK
3-to-1: AA-99, AK, AQ
2-to-1: any hand you raised with initially

Limping In:

If several players have limped in AND nobody has yet raised AND you
are in the cutoff seat or on the button, limp in with:

- Any pocket pair (66,55,44,33,22, since the orig. instructions say to
raise for 77 or higher in this situation)
- Any suited ace (A9s-A2s, since AKs-ATs = raise)
- Any 2 suited cards 10 or higher (KJs, KTs, QJs, QTs, JTs - raise
with all others, as per orig. instructions)

#Size of Opening Raise#

Overall goal: to raise the largest amount you can *while still getting
action from weaker hands*; NOT to just pick up the blinds

if you open the pot or there aren't many limpers ahead: Raise 4x-5xBb
if several players have limped in ahead, increase raise to 7.5x-10xBb
(limpers will be more likely to call since theyve already shown
interest in their hands, and you don't want to give them a correct
price to call by not raising enough since the pot is larger)

#Playing on the Flop#

"if everyone checks to you, or you are first to act, typically move
all-in"

any time your remaining stack <= 1x pot size AND you think yhour hand
might be best, moving all-in is a safe play.

In an *unraised* pot, your stack will usually be larger than the pot
size, so if you bet, bet 0.5x-1x pot size

#Summary of the MAIM#

"1) Make sure you have a small stack

2) Play very tightly. Throw away all but the best hands.

3) When you get one of the best hands, try to get all-in against a
weaker hand as quickly as possible."

"if you follow this strategy, you usually will have the best hand,
and, over time, you will make money."

I sure hope so.

So let the games begin!
I have previously tried an experiment with the Sklansky Tournament
System and Advanced System as outlined in _Tournament Poker for
Advanced Players_, but that was a horrendous failure. Neither variation
ever got me into the money in any tournament, and both systems were
completely algorithmic (no judgment involved whatsoever; you could
write a very simple computer program to implement the regular and
advanced systems). That's not to say that you can't plow pretty far
through the field with either system - you can - but as far as making
money in tournaments, it just didn't work. (Although I must note that
the buy-ins for the tournaments were relatively small ($5-$50), so
perhaps my results were tainted by a decreased resistance to the
thought of busting out of the tournament, whereas in a $5,000 buy-in
tournament, players would be much more reluctant to call an all-in bet
if it meant their busting out. Then again, usually the people who
plunk $5,000 down on a tournament buy-in have a little less inhibition
about spending money than an average online player has, so the two factors may just cancel each other out).

However the initial indications look good for the MAIM (I made $16 so far!
[img]/images/graemlins/smile.gif[/img]), but no matter which way it turns out, I'll keep you all posted.

Thanks,

Jordan

[radioheadfan]

Interested to hear the results. Please tell us what sites, what stakes, and what time of day you were playing.

If possible, track the entire experiment with pokertracker so that we can all look at the hand histories later.

Or at the very least save all of the hand histories so a 2+2er can upload them into pokertracker and run an analysis later, after you run at least 10K hands.

Best of luck!

[Jordan Olsommer]

I forgot to add, I'll only be selecting tables with 50% or more players seeing the flop, so as to control for the concern that the system might not work if the tables are too tight or good. This way, with 50%+ seeing the flop, they definitely aren't too tight, and being 0.25-0.25, they damn sure ain't too good. [img]/images/graemlins/smile.gif[/img]

Jordan

[Jordan Olsommer]

That was going to be my first question - I've played a few sessions so far, but I don't know how to get the exact hand histories out of PokerTracker - all I can seem to get it to show me are my overall statistics for each stakes level. Any ideas?

[Zag]

An interesting experiment, and I am also curious to hear the results. I'll bet you make a small profit over the long term. However, a couple of comments:

There is zero chance I could ever play this way, no matter how profitable. If I want to watch paint dry, then I'll watch the endless home decorating shows with my wife. [img]/images/graemlins/smile.gif[/img] But if it was adequately profitable, I might try hacking some of the client software so that I can make a robot that plays this way.

Rather than creating a new account, you might just have kept careful records using your normal account. Then you gain the benefit that anyone who keeps records on you will have them badly skewed if their records include these sessions of odd play.

[Jordan Olsommer]

[ QUOTE ]

There is zero chance I could ever play this way, no matter how profitable. If I want to watch paint dry, then I'll watch the endless home decorating shows with my wife. But if it was adequately profitable, I might try hacking some of the client software so that I can make a robot that plays this way.


[/ QUOTE ]

Heh, well it doesn't exactly require laser-like focus. It's more along the lines of "catch up on some reading until I hear the little beep that tells me the action's to me, and then a quick glance at what has transpired thus far. It's not playing in any sense of the word. Or to put it another way, the fun isn't in the poker (where it usually is), it's in the testing of this particular poker system to see if it really is possible to fleece large numbers of supposedly intelligent human beings in a fair manner by playing poker like a robot monkey....with big chainsaw-hands . BZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZ!

I really hope you were joking about that "making a bot" bit, though.

And about throwing people off by using an existing account, I am using an existing account (I dont think I said I created a new one, but I think I might have given off that impression with my choice of words - if so, I apologize). However, even if it were a new account, I don't want to underestimate my opponents but I really doubt that a whole lot of players are keeping a book on me at 0.25-0.25

[davedave]

shut up zag. you suck

[Zag]

[ QUOTE ]
shut up zag. you suck

[/ QUOTE ]
LOL!! I win.

[Jordan Olsommer]

Results of Day 1

Starting Bankroll: $300
Ending Bankroll: $351

Comments: Thank heavens I can do other stuff while doing this because dear god is it boring. Not to mention painful when I'm being blinded away and have to toss 66 because I'm two off the button. Also, I've learned that people don't pay attention to your "table image" at these levels. Well, I guess if you slap them in the face with it (e.g. go all-in every pot), it is conceivable they might get the drift after a half hour or so :P. But other than that, I dunno...

I have to find a way to wrangle the actual hand histories out of PokerTracker somehow so I can isolate the experiment, and the data won't be tainted by the statistics from that awful time six months ago when I tried to play $10-$20 heads up. Again, if anyone knows how to do this, please let me know.

[DavidC]

[ QUOTE ]
If I want to watch paint dry, then I'll watch the endless home decorating shows with my wife.

[/ QUOTE ]

I just wanted to quote this as something never to say to your wife. [img]/images/graemlins/smile.gif[/img]