In my ongoing quest to get my butt to play higher levels, I decided to go back and look at what my variance has really been like at the various levels I have played. I wanted to look at variance in a different way, a way that had intuitive meaning for me. And with the onset of fantasy football season, I thought that counting "wins" and "losses" seemed like a good idea.

I have all of my results in an excel spreadsheet. I have them broken down by game played ($50 NL, $100 NL, and so on), with all the ups and downs from all the days catalogued. With the magic of excel and visual basic, I managed pretty quickly to do something that I found interesting.

Let's look at $50 NL. I call it a "win" every time I get ahead by x dollars, and a "loss" every time I get behind by that same x dollars. Every time I get a "win" or a "loss," I set the dial back to zero and start a new "game."

(Note: there are limitations to my data, because I only record the amount that I won or lost when I get up from the table. I could be down $100 but then win back $75 during the same session and I would only record the $25 loss, not the fact that at the low point I was down $100. So that understates some of the variance, I guess.)

So let's say a group of sessions looks like this. Session #1 I win $42, session #2 I win $90, and session #3 I lose $100. And say that I am counting a "win" or a "loss" every time I get ahead or behind by $50. I would book two "wins" during session 2, with a net after that session of +$32 (42 + 90 = 132, subtract 100 in booking the two "wins," net =$32). Then in session 3, $32 - $100 = -$68, so I would book one "loss" during session #3, with a net after that session of -18.)

OK, it's geeky. But luckily I can let excel do all the calculations for me. It takes about two seconds, now that I wrote the program.

At $50 NL, and setting x to $50, or one buy-in, it turns out to be a pretty high-variance game. It turns out, running the program, that I am 166-54 at that game. Not a bad winning percentage, but there's definitely some frustration from losing 54 times.

Set x to $100, or two buy-ins, and things look a little better. I'm 73-17 at that game, for an 81.1% winning percentage. And setting x to $200, or four buy-ins, my all-time record is 33-5, or 86.8%.

Here are my results, for what they're worth, at the 3 games I have played the most: $20 NL (all at Prima), $50 NL, and $100 NL. "Base" means the amount I have to win or lose to record a loss at that "game."

$20 NL
Base ... Wins ... Losses ... Percentage
$20 127 21 85.8%
$40 57 4 93.4%
$60 36 1 97.3%
$80 26 0 100.0%
$100 21 0 100.0%
$120 17 0 100.0%
$140 15 0 100.0%
$160 13 0 100.0%
$180 11 0 100.0%
$200 10 0 100.0%

$50 NL
$50 166 54 75.5%
$100 73 17 81.1%
$150 43 6 87.8%
$200 33 5 86.8%
$250 25 3 89.3%
$300 21 3 87.5%
$350 18 2 90.0%
$400 15 1 93.8%
$450 12 0 100.0%
$500 11 0 100.0%

$100 NL
$100 27 10 73.0%
$200 10 2 83.3%
$300 6 0 100.0%
$400 4 0 100.0%
$500 3 0 100.0%
$600 3 0 100.0%
$700 2 0 100.0%
$800 2 0 100.0%
$900 2 0 100.0%
$1,000 1 0 100.0%

Anyway, this was a lot of fun, and very encouraging. So far I have actually had much fewer swings at $100 NL than at the other two games -- I've never lost 3 straight buy-ins at $100 NL. Obviously this is partly due to the fact that I have played the least at that level, and I can expect some bigger downswings in the future.

If anybody on the forum has their data from all of their sessions recorded this way ($ per session, and sessions in order, arranged by type of game), and if you want to send me the raw data, then I'd be happy to run the program and get you your "wins" and "losses" records. If you're willing to admit to everybody that you're interested, I will post your records publicly as well.