there is no way you'd win a lot of money if your playing 5$ sngs and i don't understand what your problem with caesar is. ill play you at my normal stakes if you want.
Notifications
Clear all
I wonder if it is possible to model variance in cash games. It probably has too many variables due to taking into consideration the many playing styles of people and using individual hand data instead of percentage of ITM finishes in tournaments but it would be good to compare with that of SnGs.*
*may not make complete sense. I'm tired.
Originally posted by ihufa
there is no way you'd win a lot of money if your playing 5$ sngs and i don't understand what your problem with caesar is. ill play you at my normal stakes if you want.
We probably played 2$ HU SNG together (I play that and PLO2 when I am really pissed) and I justed donked my stack away because they are too slow for me 
And I probably managed to suck out on vladd, so he is angry as he should be, because good players like him never suckout ♠_love: BTW. To my defence you can always check my rare HU results on SharkScope if you like:
Now something to the point as I feel like I should not waste this place with useless posts:
@davodka
I do not really know how could I simulate that as it is really depending on lots of things but the question definitely makes sense to me 
In a SSS environment it should not be that hard, but for BSS I really cant imagine anything
i have a question: if my roi is 2.5% over 175 $23 hu sngs, what is the probality that i have
1) a roi >5%
20 a losing roi
another question is my formula right for calculating roi in hu sngs.
(2n*i)-(n+r) *100%
(n+r)
n=buyin
i=itm (decimal)
r=rake (decimal)
Originally posted by AugustusCaesar
Thats how sick variance really is .. I felt a little scared when I first saw the results .. I will continue with reseatching with different ROI% (next will be 10%) and also with more players (next will be 18 player SNGs)
push !
long time ago i know. Are you going to do the math for the 18plr anytime? 
Would be nice.
thx JG
I am a computer scientist and one of my research fields is network simulations. The point of network simulation is to gain knowledge about the behaviour of a system and the performance. The point is, that the system to be modelled is mostly known to the researcher...
The problem with poker simulations is the huuuuuuge variance and standard deviation. You cannot model around this - or it does not really make sense. You can either abstract from the acutal play as done here but this does not really give you any insights. Poker is variance and a portion of luck - and last but not least - discipline...
I don't mean to badmouth these calculations but I don't rally see any significance...
I think it gives us a little insights.
Sure, variance is huge but u can do some calculations under special conditions to give us a liitle insight of how games could run if the circumstances doesnt change.
In realty it is obv more complex. Circumstances could change every month or even every day when u change ur game cause u play tired or too long sessions. These things affects ur play a lot.
However, i like these calculations because it is fun too see what could happen under unchanged conditions.
So, im looking forward to see a 18plr calculation if Augustus is still interested in this.
Greets
Bloody Geeks. Sod it, calculate this - I'm all in! Lol.
Seriously, what a great place this is and what a community we have! Experise everywhere and so willingto help out; and all free! This place is the nuts.
Wow solid thread!
I'm REALLY interested in some of this stuff about MTSNG's tho atm... They're obviously getting really popular but some numbers and figures about variance would really help tbh.
In concreto:
About true ROI over sample?
About possibility of downswings of 50 / 100 /200 BI's?
Would be nice if you could help out
thx in advance
Great blog Caeser!
It is well known that the variance is a lot higher in turbo sngs when compared to normal speed sngs.
I'm not sure how you can model this, but can you show different simulations for normal and turbo so we can actually quantify the expected variance difference between the two?