|
|
| Author |
Message |
Monica
Joined: 06 Jan 2006
Posts: 9
Location: Sweden
|
 Posted: Mon Aug 28, 2006 7:32 pm Post subject: Standard Deviation and Losses |
 |
|
Hi all,
Since I am (according to Poker tracker) losing about 1.5 BB/100 hands I have started to wonder about the probability that I am a lousy player versus being unlucky. I have played approx. 15k hands at 0.5/1.0 and my standard deviation/100 hands is 11 BB. Shouldn´t it be possible to calculate how likely it is that I am a losing player (using for exampel 95% confidence)? It was too long ago since I studied any statistic so if someone would fresh up my mind I would appreciate it a lot. Thanks! 
/Monica |
|
| Back to top of Forum |
|
 |
Bugsbunny
Wascally
Joined: 07 Apr 2004
Posts: 7752
Location: Drinking Carrot juice
|
| Posted: Wed Aug 30, 2006 6:26 am Post subject: |
|
|
That's a low SD. I can just about guarantee that you're way too passive.
Anyway I have bad news for you. Assuming:
15,000 hands
earn of -1.5BB/100
SD of 11 BB/100
Your earn range with a 95% confidence level is -3.25 to +.25
The chances that you're actually a breakeven player, or better are smaller than 5%
So, truth be told, you're probably a lousy player. That's the bad news. The good news is you can get better and this is a good place to do it. |
|
| Back to top of Forum |
|
|
the_hawk
Chelsea FTW!
Joined: 13 Jul 2005
Posts: 4452
|
| Posted: Wed Aug 30, 2006 6:52 am Post subject: |
|
|
Not for the first time, I will go for the assist... with the "How To Calculate It" bit...
1. Work out your SD for your entire sample. SD scales as the square root of the total number of hands, so a SD of 11BB/100 translates into:
11 * sqrt (15000/100) = 134.7BB for the whole sample
2. Convert this SD into BB/100; you have played 150 batches of 100 hands here so your SD corresponds to:
134.7 / 150 = 0.9 BB/100, give or take.
3. Your true earn is within 1 SD of its apparent value with 68% confidence, and within 2 SD of the mean with 95% confidence.
Hence you are 68% confident it's between -0.6 and -2.4 BB/100, and 95% confident it's between +0.3 and -3.3BB/100.
4. I won't bore you with the formula but if you have MS Excel there's a nice little function that gives you the result you want straight away. To find out the chance of being a losing player:
NORMDIST(0,-1.5,0.9,TRUE) where the three numerical arguments mean the desired earn (0 here being break-even), the mean and the SD respectively.
This tells us your chances of being a losing player, at the moment, are about 95%. |
|
| Back to top of Forum |
|
|
Monica
Joined: 06 Jan 2006
Posts: 9
Location: Sweden
|
| Posted: Thu Aug 31, 2006 6:46 am Post subject: Thanks a lot |
|
|
Thanks a lot guys. This was exactly what I was looking for. Now I "only" have left to figure out how to improve (besides being more aggressive) 
/Monica |
|
| Back to top of Forum |
|
|
Mowgli
Joined: 11 Sep 2005
Posts: 404
Location: Pergatory
|
| Posted: Sun Sep 24, 2006 5:47 am Post subject: |
|
|
| the_hawk wrote: | Not for the first time, I will go for the assist... with the "How To Calculate It" bit...
This tells us your chances of being a losing player, at the moment, are about 95%. |
Hawk that was a big help but can you go a bit further please and help me find out what Mathew was saying in his book with regards to bankroll management and fill in the bits I am missing. Thanks.
Playing full ring low limit .50/$1 worked up from .25/50
Just played 20692 hands.
Standard deviation from Poker tracker is 16.4 BB/100 hands
Win rates is -.85 BB/100
Can you show me how to calculate the range of earnings and possible rate of ruin that he showed on page 255 of his book.
For information here are the other stats
VP$P 19
PFR 4.53
SD Win 55.04
Please feel free to be brutal
 |
|
| Back to top of Forum |
|
|
Bugsbunny
Wascally
Joined: 07 Apr 2004
Posts: 7752
Location: Drinking Carrot juice
|
| Posted: Sun Sep 24, 2006 8:52 am Post subject: |
|
|
| Mowgli wrote: | | the_hawk wrote: | Not for the first time, I will go for the assist... with the "How To Calculate It" bit...
This tells us your chances of being a losing player, at the moment, are about 95%. |
Hawk that was a big help but can you go a bit further please and help me find out what Mathew was saying in his book with regards to bankroll management and fill in the bits I am missing. Thanks.
Playing full ring low limit .50/$1 worked up from .25/50
Just played 20692 hands.
Standard deviation from Poker tracker is 16.4 BB/100 hands
Win rates is -.85 BB/100
Can you show me how to calculate the range of earnings and possible rate of ruin that he showed on page 255 of his book.
For information here are the other stats
VP$P 19
PFR 4.53
SD Win 55.04
Please feel free to be brutal
|
Let's try this for starters
You can probably drop your VPIP a few notches, although that's not crucial. 19 is playable, but I suspect that there's not too many people capable of playing it correctly. Playing that many hands probably gets you in a trouble a few other ways.
However, your PFR is pathetic - especially for a VPIP of 19. That is a number you simply must get higher, whether or not you drop your VPIP. With your current VPIP you can easily double your PFR (and then some).
Earn range:
you've currently played 20,692 hands. That's 206.92 groups of 100. That is your sample size (since your units are per 100).
The Square Root of that is 14.3847141091
Your current SD = 16.4
SD/ (SQRT(sample size)) = Standard Error
In this case it's 16.4/14.4 = 1.13888888
That is Equal to 1 SE of your Earn. Now we get back to the percentages. SE has the same distribution as SD. So your results will be as follows:
+- 1 SE = 68.3%
+- 2 SE = 95.4%
+- 3 SE = 99.7%
It turns out that multiply the current SD by either 1, 2, or 3 (depemding on what % you want to get to) is a very close approximation of the actual number. so, for a desired confidence interval your true earn would be:
Observed-earn +- (SE_of_Earn * confidence_interval_factor)
For 95% that gives us:
-.85 +- (1.14 * 2) = -.85 +- 2.28
So that gives us an earn range of roughly
-3.1 to +1.4
Since you're currently in the red the chances that you're actually a losing player are obviously greater tha 50% (your current earn is the 50/50 point for any estimation along these lines, for obvious reasons) The chances that you're actually a losing player at the moment work out to about 77%, but there's still a lot of margin for error here. |
|
| Back to top of Forum |
|
|
Bugsbunny
Wascally
Joined: 07 Apr 2004
Posts: 7752
Location: Drinking Carrot juice
|
| Posted: Sun Sep 24, 2006 8:56 am Post subject: |
|
|
| Anyway pfr's of 19 require good postflop play, an area I think your probably still lacking in. Drop the VPIP a few notches, raise the PFR a LOT, and work on your postflop game. |
|
| Back to top of Forum |
|
|
Mowgli
Joined: 11 Sep 2005
Posts: 404
Location: Pergatory
|
| Posted: Sun Sep 24, 2006 1:09 pm Post subject: |
|
|
OK Bugs,
Thanks for your response. I have taken all the advice onboard. Just as background, I have based my starting hands on a tighter version of the charts in ITH. I shall review them after this post and tighten up.
I do understand that once you have your pre-flop strategy in place then it is time to move on and really get to grips with the post flop game. To get a better understanding, I have read and reread and try to apply the advice given in SSH and HPFAP by Sklansky. However on checking my hands I have identified a number of leaks thant need to be plugged both pre and post flop.
Out of curiosity, is there any software or spreadsheets out there that would save me working out the equations you have shown? |
|
| Back to top of Forum |
|
|
Bugsbunny
Wascally
Joined: 07 Apr 2004
Posts: 7752
Location: Drinking Carrot juice
|
| Posted: Sun Sep 24, 2006 6:59 pm Post subject: |
|
|
| Mowgli wrote: |
Out of curiosity, is there any software or spreadsheets out there that would save me working out the equations you have shown? |
Since you asked
(see links in post below)
Should be fairly self explanatory. Any questions ask here. The worksheet is protected so you can only enter data in the changable cells, but you can unprotect it if you want access to the other cells.
Last edited by Bugsbunny on Wed Sep 27, 2006 12:21 pm; edited 1 time in total |
|
| Back to top of Forum |
|
|
janeg
Regina Canada
Joined: 04 Oct 2004
Posts: 5241
Location: Somewhere down the crazy river
|
| Posted: Sun Sep 24, 2006 10:02 pm Post subject: |
|
|
| Thanks for the spreadsheet Bugs, much better than the one I was trying to work out by myself |
|
| Back to top of Forum |
|
|
Mowgli
Joined: 11 Sep 2005
Posts: 404
Location: Pergatory
|
| Posted: Mon Sep 25, 2006 1:04 am Post subject: |
|
|
Bugs been to the link but the file does not appear to be there any longer. Would u mind uploading it again.
Thanks |
|
| Back to top of Forum |
|
|
Bugsbunny
Wascally
Joined: 07 Apr 2004
Posts: 7752
Location: Drinking Carrot juice
|
| Posted: Mon Sep 25, 2006 3:19 am Post subject: |
|
|
| Mowgli wrote: | Bugs been to the link but the file does not appear to be there any longer. Would u mind uploading it again.
Thanks |
Try this. And try using IE if you normally use Firefox.
http://dl3.ohshare.com/v/5606529/HandRangeEV.xls.html
You should just be able to click on where it says "Download file", but for some reason the link doesn't work in Firefox (at least my copy)
Alternate location. This one has a 15 second delay
http://www.mysharefile.com/v/6362897/HandRangeEV.xls.html
If that doesn't work let me know and I'll find another site. |
|
| Back to top of Forum |
|
|
Mowgli
Joined: 11 Sep 2005
Posts: 404
Location: Pergatory
|
| Posted: Mon Sep 25, 2006 4:55 am Post subject: |
|
|
OK Bugs - got it. Cheers.
On an aside, your comments made me really go to town on Poker Tracker and my hand histories.
I have identified many, many mistakes and some of my positional play as been nothing short of attrocious.
In my arrogance I thought I knew what I was doing because I had read a few books and was just unlucky.
I needed a kick.
Will post again in a week. |
|
| Back to top of Forum |
|
|
AlamedaMike
2K Club
Joined: 29 Jun 2005
Posts: 2042
Location: Alameda, CA
|
| Posted: Tue Nov 28, 2006 1:35 pm Post subject: |
|
|
Interesting spread sheet, Buggs, thanks for sharing - I have always had a curiosity in stats and I have forgotten most of it anyway.
I was wondering if my rate per hour followed the normal distribution so I did some calculations. I have 229 6/12 sessions recorded in Statking for this year and I put them in spread sheet.
I don't know how to post a chart - I should learn - in any event. The numbers below show the number of observations in the range. I did the ranges by hand so I used $30/hr ranges.
You can see that 49 obs fell between -30 and 30. It might be a little smoother if I use $12 ranges - 1 BB.
It looks pretty standard to me given 229 observations.
Obs Lower Upper
4 -270 -241
4 -240 -211
9 -210 -181
7 -180 -151
13 -150 -121
23 -120 -91
19 -90 -61
23 -60 -31
49 -30 30
19 31 60
19 61 90
11 91 120
4 121 150
3 151 180
6 181 210
6 211 240
2 241 270
edit - BTW - I figured out how to do a frequency matrix so I do not need to do it by hand. |
|
| Back to top of Forum |
|
|
grignard
Joined: 16 Aug 2004
Posts: 184
|
| Posted: Mon Jan 15, 2007 6:48 am Post subject: |
|
|
Would anyone still have a copy of this spreadsheet? I'd be interested in looking at it but the link(s) aren't working correctly for me.
Thanks. |
|
| Back to top of Forum |
|
|
|
