Why 20 buy-ins is not enough

I haven't posted in a long time, and I won't post again for a long time.

I'll post this to make a point: if you're averagely good, 20 buy-ins cannot be considered safe bankroll management. If you do the math, you'll see that over this stretch of 419,000 hands I ran at 1.46ptBB/100. And I still pulled off two swings that would make the devil cry.

Please go ahead and attribute all (or most of this) to tilt, if it makes you feel better. I used to do that, too, when someone posted a sick swing.
However, having looked over these swings in a lot of detail, I'd give the following numbers:
Swing 1 (-36 buy-ins): about 30%-40% attributable to tilt
Swing 2 (-50 buy-ins): about 10% attributable to tilt

Take it or ignore it.

P.S.: Stakes are all 6-max and some HU, buy-ins $0.25/$0.50, $0.50/$1, $1/$2, $2/$4, $3/$6, $5/$10, $10/$20

Attached Images

notenough.jpg (90.5 KB, 50 views)

This is beyond my comfort zone but that graph shows you make a valid point.

yeah i have always set up a standard roll amount as limits go up.

10nl-25nl = 20 BIs

50nl= 30 BIs

100NL=35 BIs

200nl= 40 BIs

I think passed 200nl 40-50BIs is a good number, especially since these levels are a lot of the time used for income and better safe then losing your house lol.

hmm I think I may wait till I've got 30 buyins till I move up to 100NL

Good thing you didn't post the $ graph. I guess that's even scarier...

And I agree, beyond micro stakes, 20 buy-ins is nothing.

oh and why?

are u still playing chris ?

Yes, playing the micros.

Chris is you don't mind, if you use HEM, could you let us know your Standard Deviation?

66.94 big blinds. What does that mean?

it's what you put in SD box here:

http://www.pokervariancesimulator.fr

pretty sure it means you're a deranged maniac

I like this post very much so

...me too, and thanks for the link belgo. Interesting to think about, you could play 100k's of ev+ poker and still lose if you play a high variance style. Puts downswings into perspective.

fyp

True, I was looking at 2BB+/100 winrates which would be very unlikely to lose over such a sample playing low variance, but still possible. Lower winrates which would be good at midstakes are obv. more prone to long breakeven or losing stretches.

i'm still missing something. if you dont have 20 buyins for the level your on you would drop down. yes you can lose more than 20buyins do to variance and not bad play. but this is why we put BR management into action. allowing for more buyins is ok. but i think 20 would be good enough to start a level. obv you cant stay at the level and lose all 20 buyins. you would need to have 20buyins at the level to stay at that level. and 20 buyins for the next level to move up.

sd.jpg (http://www.cardschat.com/attachment.php?attachmentid=22543&d=1249042320)
Wat does this tell me?

stop playing poker? lol wtf happened, glitch?

Avoid hands that are yellow or teal.

It reinforces my point. Out of 100 players with a true winrate of 1.65ptBB, one player managed to be stuck around 50 buy-ins after 225,000 hands. If that's not scary, then I don't know what is.

this.

Also, you can't really know your true winrate until you play something like a million hands.

I can't be bothered to create and attach an image of my graph, but over this year I'm at 3+bb/100 for 200NL, and I've had one 30+BI downswing and another 20+ BI downswing. So yes; if moving down in stakes is not an option (maybe because you need to play the limits you do in order to support yourself and/or your family) then 20 buy-ins is definitely not enough. If I absolutely could not afford to move down and wasn't playing fulltime for a limited time only, I'd require a 100 BI bankroll to play 200NL.

This calculator is really great for estimating your winrate.
http://www.castrovalva.com/~la/win.htm

If I plugin my own data and want 99% confidence, then my true winrate is between 0.13ptBB and 2.79ptBB.
Even assuming one million hands, a 99% confidence gives a fairly wide ranged 0.6ptBB to 2.32ptBB.

So imagine this: you ran at 1.46ptBB over one million hands, and there's a larger than 1% chance* that you ran more than twice as well as you should have. Over the course of one million hands. Talk about a hot "streak".
*someone know the actual number, I can't do that math?

Variance in poker is so much larger than our puny little minds could ever imagine.

i want to shoot variance with a bazooka

hu variance is the best

Can someone break this information down for someone who is poker tracker inept,such as me.Thanks

Step 1: Imagine a really big bankroll for whatever stakes you play.

Step 2: Realize that you should probably have at least twice whatever you just imagined as your real bankroll.

Ok so if i think 50 bi is plenty it should be 100bi's? If this is right why may i ask.

Indeed, Ed Miller (et. al.) conclude (the very last page) their recent e-book with this:

br 100x.jpg (http://www.cardschat.com/attachment.php?attachmentid=22575&d=1249237998)

Thanks alien i see the point but if your not playing for a living,an you have 50 bi lose 10bi's an drop down thats fine brm correct?

If you're not playing for a living, very little of bankroll management applies to you at all. And I don't mean for that to sound as facetious as it probably does, but it's a fact - bankroll management for recreational players is mostly a matter of keeping the games fun to play. For purely recreational players (typically found at 50NL and below, if we're talking no-limit, but stakes may vary) bankroll management is an exercise in playing with discipline; useful for later on but not technically necessary. For many semi-professional players (typically play 100NL+, but have another main source of income) bankroll management is mostly about not having to drop down when they come across a tough stretch. The reason they don't want to drop down is because it hurts their ego a bit (I would know; I'm the same).

For professional players, having to move down can mean the difference between paying your bills and not paying your bills because having to move down will cut your expected income by a lot. And on top of that you have the psychological problems of actually having to worry about your bankroll when you should focus on playing well.

Now, that said, I practised fairly heavy discipline with my bankroll starting the minute I learned what bankroll management was all about. Now that I live off of my poker earnings, however, I also appreciate that when I played recreationally, my bankroll discipline had more to do with pride than it did with security. I could theoretically have played every level with a very short bankroll since moving down in stakes wouldn't have had (not by much, anyway) any meaningful effect in my monthly income. The psychological reason I didn't was that I considered moving down "failing." The practical reason I didn't was because by forcing myself to play until I had doubled an already substantial bankroll was to make sure I was actually beating the game before I moved up to the next level.

For SnG's would this be good Bankroll management. When you get 100 buy ins +1 of the next level move up. for instance if you have $99 you should play $1 tournaments. but when i get to $113.40 move up to the $3.40 Sng's then when i move up to $345.50 move up to $5.50. if you drop back below the threshold then move back down again?

that seems really conservative

i think 50 buy ins would be appropriate.

if you're talking about turbos, closer to 100 is probably good