Quick question: I understand that spins are really high variance and have played with swongsim quite a bit and feel like I'm beginning to grasp this concept decently.

What I'm not clear on, is it just the wins and payouts that are high variance, or is cEV itself high variance. Do winning players go on long losing and break even stretches, but their cEV line stays consistent or is the cEV line also prone to high variance?

if you use cev mode on swongsim, its on a dropdown on the graph and use a standard dev of i think 250(?) it might be 350 or 450 iv not done it in a very long time, hopefully somebody here can correct me, you can see the swings you will get in cev there...

quick answer is yes, cev variance is very high

i dont know if what i will state is scientifically correct but from my personal experience (around 30k games) on mid stakes 30-60 i havent seen a big variation on cev when the sample is close to 1k games.

It's 400 over my last 1928 games at Stars/Party (ABI $0.63), perhaps too low because of my overly longball style (I find myself in more all-in races per game than a good reg would). Max Cut once calculated the standard dev of 440 over 1748 WPN jackpot SnGs.

Probably best to pretend as if CEV variance isn't very high, especially if you're just beginning! In any case at <$10 BI where you should be getting 80+ CEV, CEV variance should not be all that high (as in, over a 1k sample, you should be able to notice a trend).

Thanks for all the replies. I have 76cev over about 700 spins and was just trying to figure out how accurate that is. I know the sample is small but I can only put in like 100-200 spins a week right now, so I feel like it's taking forever to get a relevant sample.

Sent from my Pixel 2 using Tapatalk

I don't mean to steal your post, but I was wondering the very same thing about the degree of variance in spin'n'goes. It is obviously huge! My first 500 tournys, have a chipEV of 79 per tourny, but actual chip winnings are all the way down at 25 per tourny (lucky me won a 25-multiplier in there to keep net$win positive still). My first goal is to play 10000 spin'n'gos before deciding to continue or go back to cashgames, but is this what I have to get used to or am I running on the bad end of the scale here?

pretty standard, but the op question was about chip ev variance, not about chip variance

Aha, now I see! Thanks for clarifying. Also a very interesting topic indeed that I will keep in mind going forward.

i think ChipEV variance is the lowest one of them all, followed by EV adjusted results, then actual results being the one with most variance.

when i look at my ChipEV graph there is almost no downswings, only breakeven stretches of around 200-400 games.

My ChipEV graph:



the actual results graph will follow the ChipEV graph somewhat, so if your ChipEV graph is good, then your actual results will also be good.

this means just look at your ChipEV graph every 500 games or so and you'll have a good idea whether you're a winning player or not. personally i just set my filter to only show the ChipEV graph and look at that at the end of every session, if ChipEV is good then i'm happy.

My ChipEV vs actual results:



My ChipEV vs all-in adjusted results:




what you'll notice is that the EV adjusted results graph kind of looks like the ChipEV graph, i mean you can spot the same sort of patterns, the upswings, the downswings, etc. however, when comparing with the actual results with ChipEV the pattern is a lot more distorted, because of much higher variance.

Isn't 4k games a fairly small sample in spins?

its more than enough volume to illustrate that cev have relatively low variance.