Originally posted by nomorefish
Hi weazel,
do you prefer any changes in RFI and CC ranges for ZOOM microstakes? thx for answer smileNo, these should be good for zoom. They were tested in zoom games.
I don't mind playing tighter cold-calling ranges at first however. I prefer to think of these ranges as the maximum that can be profitably defended. So it wouldn't be incorrect to cold-call tighter and might make sense depending on your level of experience.
may help you @Mon3yMaChin3
Originally posted by Mon3yMaChin3
Can we also use this chart and for zoom 6max?Or need be more aggresive
Certainly usable for Zoom. No need to make changes there.
Hello,
fish playing random preflop
will follow this thread and looking forward to homework ♦
Hi w34z3l ♦,
This is my homework:
1.
2.
Win rate in big blinds / 100 hands: 1.60 over 50k sample
Standard deviation in big blinds / 100 hands: 90 ♠_o:
BR: 250
6max nolimit: nl5
ROR formula I found :
ROR = e ^ (-2WB / (S ^ 2)) = 13.87
e = constant (2.718281828)
W = win rate bb/100(1.60)
S = standard deviation bb/100(90)
B = bankroll in bbs(5000)
ROR
500 big blinds 82.08%
1000 big blinds 67.36%
2000 big blinds 45.38%
3000 big blinds 30.57%
4000 big blinds 20.59%
5000 big blinds 13.87% 
Minimum bankroll
(for less than 5% risk of ruin) 7583 big blinds 
I used an online calculator from http://www.pokerolymp.com/articles/show/keine/13048/variance-calculator
My actual BRM:
........ UP ......... DOWN
nl2....25BI........ -
nl5....25BI........20BI to nl2
nl10..30BI........20BI to nl5
nl25..40BI........30BI to nl10
nl50..50BI........40BI to nl25
According to my ROR calculation I should have a more nitty BRM for a 5% ROR but I would rather reload than have a 75BI BRM.
3. ?(This is my HUD right now:
VPIP / PFR / 3BET / FOLD TO 3BET / SQUEEZE
POSTFLOP AGG / FLOP CBET / FLOP FOLD TO CBET / TURN CBET / TURN FOLD TO CBET / WTSD% (when saw flop)
EP PFR / MP PFR / CO PFR / BTN PFR / SB PFR
BTN STEAL / CO STEAL / TOT FOLD TO STEAL / BTN REST vs CO STEAL / BB FOLD vs SB STEAL
RIVER BET / RIVER FOLD 2 BET / RIVER CALL 2 BET
Great series! Keep up the good work! ♠_cool:
Hey w34z3l,
When does the next video comes out ♦ I can't wait to see it
Greeting tilt 
Originally posted by bluedream23
Hi w34z3l ♦,This is my homework:
1.
2.
Win rate in big blinds / 100 hands: 1.60 over 50k sample
Standard deviation in big blinds / 100 hands: 90 ♠_o:
BR: 250
6max nolimit: nl5ROR formula I found :
ROR = e ^ (-2WB / (S ^ 2)) = 13.87
e = constant (2.718281828)
W = win rate bb/100(1.60)
S = standard deviation bb/100(90)
B = bankroll in bbs(5000)ROR
500 big blinds 82.08%
1000 big blinds 67.36%
2000 big blinds 45.38%
3000 big blinds 30.57%
4000 big blinds 20.59%
5000 big blinds 13.87%
Minimum bankroll
(for less than 5% risk of ruin) 7583 big blinds
I used an online calculator from http://www.pokerolymp.com/articles/show/keine/13048/variance-calculatorMy actual BRM:
........ UP ......... DOWN
nl2....25BI........ -
nl5....25BI........20BI to nl2
nl10..30BI........20BI to nl5
nl25..40BI........30BI to nl10
nl50..50BI........40BI to nl25According to my ROR calculation I should have a more nitty BRM for a 5% ROR but I would rather reload than have a 75BI BRM.
3. ?(This is my HUD right now:
VPIP / PFR / 3BET / FOLD TO 3BET / SQUEEZE
POSTFLOP AGG / FLOP CBET / FLOP FOLD TO CBET / TURN CBET / TURN FOLD TO CBET / WTSD% (when saw flop)
EP PFR / MP PFR / CO PFR / BTN PFR / SB PFR
BTN STEAL / CO STEAL / TOT FOLD TO STEAL / BTN REST vs CO STEAL / BB FOLD vs SB STEAL
RIVER BET / RIVER FOLD 2 BET / RIVER CALL 2 BETGreat series! Keep up the good work! ♠_cool:
hmm isnt your HUD a bit too big? if so make small popup just for couple stats like me:
other small popup
i prefer my own popups because i know where is what. I dont like big and full of all info together. Maybe its just me. But this is how you would start use more popups. People tend to overpower huds and aint using popups at all.
Well, I'm used to it, but you got one real good point there, I don't use popups that much. It's on my "to do list" to make some popups of my own, but I didn't managed to do so yet.
Originally posted by bluedream23
Hi w34z3l ♦,3. ?(This is my HUD right now:
VPIP / PFR / 3BET / FOLD TO 3BET / SQUEEZE
POSTFLOP AGG / FLOP CBET / FLOP FOLD TO CBET / TURN CBET / TURN FOLD TO CBET / WTSD% (when saw flop)
EP PFR / MP PFR / CO PFR / BTN PFR / SB PFR
BTN STEAL / CO STEAL / TOT FOLD TO STEAL / BTN REST vs CO STEAL / BB FOLD vs SB STEAL
RIVER BET / RIVER FOLD 2 BET / RIVER CALL 2 BETGreat series! Keep up the good work! ♠_cool:
I guess btn steal and co steal is useless here?
Originally posted by AlvisR
hmm isnt your HUD a bit too big? if so make small popup just for couple stats like me:
other small popup
i prefer my own popups because i know where is what. I dont like big and full of all info together. Maybe its just me. But this is how you would start use more popups. People tend to overpower huds and aint using popups at all.
This is an excellent suggestion, AlvisR
I spent a lot of time on my popups, where each stat is listed by street and so on. The problem is that it is hard to find exactly what I want to know in the second or two (or less) I have available.
Your idea of very small, popups is a very good one.
Best of luck,
--VS
Aloha chaps. Laptop has been out of action so only just managed to get back to silver status to take part (hope rest of vids are silver...)
I have a question about winrates etc...
I tend to put in much more theoretical work than actual time playing poker (family commitments) so I think my ability is above the limits I'm playing currently. I managed to beat NL4 in 10k hands and I've moved up to NL10 and so far going well. Enough boasting, my question is regarding variance. I understand that we are talking about a SD of ~80BB/100 and comparatively (to this) low winrates of ~5BB/100 which is what causes such a variable result over a small sample size. My winrates over NL2 and NL4 have been ~20BB/100 so far so how many hands should I play before deciding I'm not on an upswing and that I'm actually beating a limit?
I've read about 100k break-even stretches but surely the chance of this is much reduced if you are crushing a limit. It would suck to think I've been on a 50k heater that is about to burn out...
Basically, I'm saying that I'm too lazy to work out a formula that expresses the distribution of expected total winnings for a given winrate and standard deviation over a given number of hands, thus allowing calculation of the likelihood that you are actually beating a limit in a given number of hands.
I know it involves a Gaussian distribution. Probably.
Anyone got any thoughts?
Basically, I'm saying that I'm too lazy to work out a formula that expresses the distribution of expected total winnings for a given winrate and standard deviation over a given number of hands, thus allowing calculation of the likelihood that you are actually beating a limit in a given number of hands.
I know it involves a Gaussian distribution. Probably.
Anyone got any thoughts?
Hi everyone, hi w34z3l,
Long-time lurker here but I thought I'd jump in and contribute as maths is something I know a thing or two about!
This is not too difficult actually. In fancy language our "bb/100" result over a certain number of hands is called a "sample statistic" and we're talking about the difference between the underlying distribution (our "true" bb/100 and standard deviation) and the distribution of the sample statistic (i.e. the mean and standard deviation of our actual results over a certain sample size). The relevant facts are:
1) the sample mean is the same as the true mean. This one is pretty obvious. If our true average winrate is +10bb/100, then on average we make +10bb/100 over any sample size.
2) the standard deviation of the sample mean improves (which means, becomes smaller) with the square root of the sample size. So if our standard deviation is 80bb/100, the standard deviation of our "bb/100" result over a 6400 hand sample (64 times as big) will be 10bb/100 (i.e. 8 times smaller).
3) as an approximation, 95% of the time an observation falls within 2 standard deviations either side of the mean.
As an example, suppose a person wins at +10bb/100 with a standard deviation of 80bb/100.
Over 100 hands, their observed winrate will be between -150bb/100 and +170bb/100 (i.e. 2 standard deviations above or below the mean) 95% of the time.
Over 6400 hands, their "bb/100" will still average +10bb/100 but with an SD of 10bb/100 (being 8 times smaller because this sample is 64 times bigger). So 95% of the time they will achieve between -10bb/100 to +10bb/100 over a 6400 hand sample.
With 25600 hands, the SD drops to 5bb/100 (the sample is 256 times bigger, square root of that is 16, so the standard deviation improves from 80 down to 80/16=5). The 95% range is from 0bb/100 to +20bb/100. So this is the smallest sample size for which this person would unlucky to have a worse than break-even stretch.
My winrates over NL2 and NL4 have been ~20BB/100 so far so how many hands should I play before deciding I'm not on an upswing and that I'm actually beating a limit?
On the face of it, 6400 hands would start to mean something. (Underlying SD is 80bb/100 which means the SD of the average bb/100 over a 6400-hand sample is 10bb/100. This means your actual result of +20bb/100 is two SD above zero.) Really important note: this does not mean that after 6400 hands you can be confident you are a +20bb/100 crushing player! It means that, if you achieve +20bb/100 over a 6400 hand sample with an SD of 80bb/100, you can just barely be confident that your true winrate is above zero at all!
No offence but it's probably unlikely you will maintain +20bb/100 over a bigger sample. Also, because of the square root thing, with lower winrates you will need many more hands in order to be confident that you are a winner.
+20bb/100 -> 6400 hands needed
+10bb/100 -> 25,600 hands needed
+5bb/100 -> 102,400 hands needed!
I've read about 100k break-even stretches
With a standard deviation of 80bb/100, a person's winrate over 100k hands would have a standard deviation of roughly 2.5 bb/100 (divide 80 by square root of 1000 to get approx 2.5). So yes, unless the person is crushing a lot more than +5bb/100, it's perfectly possible even for a winning player to break even or slightly lose over a 100k stretch.
Final comment, this is all a bit theoretical. In practice we'll never know our "true" winrate. It takes so long to get a big enough sample, that our skill and the skill of our player pool will have changed by then! So it is an impossible moving target. It's best to move up the limits based on a BRM plan and taking shots, not on waiting for statistical significance over a gazillion-hand sample.
Hope this helps!
hey weasel, great 2nd vid,
I have a question regarding the 3betting ranges though
In BB vs BTN the following range is included neither
in the coldcall range you gave in part 1 nor in the 3bet range:
ATs-A9s,KJs+,AJo-ATo,KJo+
could you give your reasoning behind this?
Thankyou so much @member001 - amazeballs answer. It's the fact that SD scales as the squareroot of the sample size that I was missing. It all makes sense now. No offence taken on the old 20bb/100, I realise that this is unrealistic and that's what got me thinking about the chance that I may not be a winning player longterm. I agree though that all we can do is move up and down as BRM dictates, play well and work on our games. Winrate is a function of skill and opponents, both of which are variable over the sample.
Thanks again!
Originally posted by Krad85
hey weasel, great 2nd vid,
I have a question regarding the 3betting ranges thoughIn BB vs BTN the following range is included neither
in the coldcall range you gave in part 1 nor in the 3bet range:
ATs-A9s,KJs+,AJo-ATo,KJo+could you give your reasoning behind this?
you mean this?
Originally posted by Krad85
yes, this "gap" between CC and value 3bet range...
AKs vs EP and MP the same
as i understand AKs hand we can do both ways successfuly. 3betting or coldcalling would be fine depending from situation. I tend more on 3betting
