www.dropbox.com/s/ukpadqy24se12th/SwongSim.exe?dl=0

• Added an option to draw fewer simulation lines in the visualizer. This speeds up the rendering time when drawing large numbers of simulations. The full number of simulations are still run and reflected in the tick marks, traces, and text output. The option is located on the visualizer tab.
  
  

I have two custom stats;

Standard Deviation of Luck Adjusted Winnings/Hand in Game Currency

Standard Deviation of Luck Adjusted Winnings/Hand in BB per 100

Respectivly 125 and 40 value... i guess i need standard deviation of chip win by game non luck adjusted right?

But it is kind of nonsense if i want look variance of EV, i need the standard deviation of EV , no of winnings...

We want standard deviation of luck-adjusted chips won / game.


I have a clunky stat that will give the sum of squared deviations (part way to the standard deviation), which I can then manually divide by the number of games and take the square root of to get the sample standard deviation. I'm not great at PT4 stat building, but if I come up with a more acceptable version I will post it.

For my small sample of 1748 games, I have a standard deviation of 442 on an average of 60 EV chips per game (median is 2.5). I made a histogram of my EV chips won for each game (below) to see how normal it looks. It appears to have some skew, although I don't know how significant it is for that sample size.

Fighting a cold atm, so apologies if I'm slow to reply over the next few days.

That's the best you can do in PT4, afaik, because tournament reports can't include this kind of stats. (Per-tourney reports group hands by tourney - there's a GROUP BY tourney_summary.id_tourney clause in the corresponding internal SQL query - so stats that summarise per-hand DB entries in a non-standard way can't be used.) I made an 'all hands report' and aggregated the squares of net chip winnings in every hand like you.

I also agree that the cEV is skewed. The 'EV difference' (between the cEV and the actual chip winnings) is not very skewed because e.g. 70/30 flips in Hero's favour happen about as often as 30/70 flips in opponent's favour for about the same pot size. But the actual chip winnings are somewhat skewed positively because they're either +1000 or -500 for a single tourney. Hence the sum of the two stats - the cEV - is skewed positively.

Get well soon!

I got something between 350 and 400 (don't remember exactly) for iPoker Twister cEV, but that's because it has big antes in the second level and is thus more conducive to coinflips. For WPN's Jackpot Poker that has a healthy 15/30 level, your SD looks plausible. Thanks for sharing it!

That makes a lot of sense. Many thanks!

HU calculations are not working.

Thanks for reporting this. Please check on the "About" tab to see if you are using the latest version (Version 1.16.5853.28481). If not, please let me know if you see the same problem with the latest version. If it is happening with the latest version, does it happen if you just start the program and select the HU preset structure? Otherwise, if you can describe the steps that lead to that state, I will be better able to diagnose the bug.

--Max

This has been an issue before, and I believe it's a bug related to the decimal in the buyin (1.95). Try change the "." with a ",".

Yeah, the '.' decimal separator needs to be replaced by a ',' if the ',' is the default separator in the 'local standards' settings of your Windows (e.g. if you live in an Eastern European country and haven't changed those settings).

^ http://forumserver.twoplustwo.com/45.../#post47153055
http://forumserver.twoplustwo.com/45.../#post49042635

Oops, disregard this, I was going to post those links into the Spin & Go discussion thread but had this thread open as well and submitted the post in the wrong browser tab Thank you for giving a detailed answer to SiQ there

Was glad to re-read that post of yours.

Max Cut,
I'm not sure if it's been mentioned (or maybe it's already an option that I can't find) but having a log-log graph option for the visualizer would be really nice.

Yeah, something should be done to control the scaling for those very large multipliers. Thanks for the good idea; I have added it to the to-do list.

Example of the problem where a graph is almost useless because of the high multipliers:





I took a peek at what log-log and then semi-log (log on y-axis) would look like and was not pleased with them for what is supposed to be an intuitive visual aid.

log-log:





Semi-log:





So, for now at least, I just made it possible to draw a larger graph that can be scrolled.

• Added option to specify the vertical size for drawing graphs. The option is located on the visualizer tab. (Mouse-wheel zooming is now a bit wonky due to the addition of scrolling, but it's on my list to fix.)

Download link:
www.dropbox.com/s/ukpadqy24se12th/SwongSim.exe?dl=0


-Max

Would be really good if you can add an option for show EVroi% instead of chip/game in all the EV chip mode simulations
Most of the NONspins regs use evroi instead of chip/game!

thanks a lot! and thanks again for your work!

Edit; and i still looking for custom PT4 stat of standard deviation of luck-adjusted chips won / game
I didnt fully understand how to reproduce the result your last post about it

ty!

Thanks for the suggestion. I agree it could be convenient. Unfortunately, the relation between EV roi% and EV chips/game is not the same for all formats. Added to list to think about.

I am also still looking for an easy way to determine standard deviation for EV Chips won per game. Essentially, what I did was add all-in adjusted equity to my tournament results report, export the report to a csv file, and then import it to a spread sheet program so I could manually calculate the standard deviation. If I find something easier, I will be sure to post it.

Also, just a reminder that the chip mode simulations are not a perfect representation since they do not include the skew seen in actual EV chip distributions, so I advise caution when interpreting them. Still some work to do on this feature.

-Max


Download link:
www.dropbox.com/s/ukpadqy24se12th/SwongSim.exe?dl=0

I think (not sure) you could have played around with the factor to get a better looking graph using log-log, but I actually really love your solution so I'll just shut up and enjoy it awesome job with that.

Hi.

First of all, thank you very much for this amazing tool.

I´ve realized that you use 5% as the rake in this modality. That´s not true. For example in 100$ spin and go is 95 +5. So the rake is 5/95 = 5.263 %

Another trick of pokerstars :/

That´s quite relevant in this game. When you make the calcs to get the rake with the total prizepool would be :

(1000000*3 - 2850000) / 2850000 = 5.26
(1000000*3 - 2814000) / 2814000 = 6.61 removing prizepool for 1º prize
(1000000*3 - 2802000) / 2802000 = 7.07 removing prizepool for 1º,2º prize
(1000000*3 - 2790000) / 2790000 = 7.53 removing prizepool for 1º,2º,3º prize

Do you agree with that?

Thanks for your time

It's up to the terminology. Before year ~2014, rake used to be calculated as a percentage of the 'pure' buy-in (buy-in minus rake) because most of the buy-ins were non-round (like $100+$9). But nowadays, poker rooms like to make the total buy-in 'round' ($100 looks prettier than $109) and $100 is easier to divide by than $109. So it's now standard to calculate and compare the rake as a percentage of the total buy-in, i.e. $5 / $100 in our case, not $5 / $95.

Yep, it's just an accounting detail and changing it would have no affect on swings, EV ROI, etc. This format is chosen to reduce confusion by being in agreement with numbers used on the websites.

Cheers
-Max

Download link:
www.dropbox.com/s/ukpadqy24se12th/SwongSim.exe?dl=0

Yeah i was wrong.

Thanks guys.

i think i find one bug, (or it s my mistake)
For HuSNG
In my pt4 3% roi = 36 chips / game
But in the evchip window, when i set 2 players 36chip/games win
Results are 5%roi
I set the rake correctly (2%)


thanks

Hmmm... win rate looks correct.
(500 + 36) / 1000 = 0.536

ROI looks correct.
0.536 x 1.96 - 1 = 0.0506

But I've been awake for a long time so maybe I'm missing something. Is it a custom PT4 stat?

-Max

Download link:
www.dropbox.com/s/ukpadqy24se12th/SwongSim.exe?dl=0

That's likely because you've played a mix of stakes and your chip EV is lower at higher stakes*. Try filtering to include one only one stake at a time.

* E.g. if one mixes 200s, 300s and 500s in the 1:1:1 proportion, wins 56 chips at 200s, 36 chips at 300s but 16 chips at 500s, then the chip EV is (56+36+16)/3 = 36 per game, but the ROI is (9% * $200 + 5.1% * $300 + 1.1% * $500) / ($200+$300+$500) ~ 3.9%. The difference between your chip EVs at different stakes is likely even bigger because the sample is small and you just ran the worst at your highest stakes in terms of the EV.

That's why Sharkscope offers the average ROI stat (which is 5.1% in the above example, corresponding to 36 chips per game) instead of the total ROI (3.9%). The former is a less precise indicator of ability in the long run (because higher stakes are tougher) but it's subject to smaller swings than the latter.