Tell me what you think

[sagefr0g]

Originally Posted by Kasi View Post
Neither can change the laws of probability.

True -- but legitimate forms of advantage play take advantage of those laws to produce favorable outcomes. Betting systems do not.

there is a question on this laws of probability stuff that leaves me clueless.

like, ok i know there is the law of large numbers. i know it's pretty much what the casinos have working for them when they offer negative ev games and it's pretty much i guess what counters have working for them when they make a negative ev game a positive one. then i guess a counter strives for his N0 and maybe that N0 in a sense gets him into the 'realm' of the law of large numbers, sort of thing.
but anyway, the question i wonder about (and i've asked a similar question before receiving basically two opposing answers, lol) has to do with variance and expectation.

i think the original question i asked was, say a guy has been playing blackjack as a counter against some game conditions and has experienced one heck of a lot of luck, to where his results are way higher than his expectation. so does that mean one is 'due' some really bad results to even up the score? i guess the answer was yes and no, lol. yes, your going to go 'down hill' some maybe as your results will approach expectation, but no, it doesn't have to be in hugely disastrous sort of way (but maybe it could be).
like the descent to results meeting expectation could go in a long leisurely albeit 'down hill' ride, lol.

so anyway, now i guess more what i'm wondering, is to put it crudely, does the laws of probability keep a 'rap sheet' on players, lmao.
sort of like the question of the statement, of how blackjack is just one big long game, sort of thing.
i mean ok, say a guy plays blackjack all sorts of ways. maybe flat bets just basic strategy at times, then counts and bets 'orthodox' at times.
well, say while flat betting one experiences a lot of luck and wins a load of money, way above the expectation of a perfect counter. then he turns around and puts in some 'serious orthodox' counting sessions.
question being, is the guy now playing as a serious counter going to 'pay' for his flat betting luck sessions by losing a load of money to get him statistically where he should be?:whip:

 

[Sonny]

i think the original question i asked was, say a guy has been playing blackjack as a counter against some game conditions and has experienced one heck of a lot of luck, to where his results are way higher than his expectation. so does that mean one is 'due' some really bad results to even up the score?

Ah, the Gambler's Fallacy rears it's head again. The player is not "due" for anything other than his EV (plus or minus the variance). Just like seeing the roulette wheel land on black 5 times in a row does not mean that red is due, having many lucky sessions does not mean that an unlucky one is due (although it often feels like that happens). The results can "even out" even if you never have another losing session for the rest of your life. Here's an explanation from Wong's site:

http://www.bj21.com/boards/free/free_board/index.cgi?read=153264

Here are some similar ones from this forum:

http://www.blackjackinfo.com/bb/showthread.php?p=67906
http://www.blackjackinfo.com/bb/showthread.php?p=111135

i mean ok, say a guy plays blackjack all sorts of ways. maybe flat bets just basic strategy at times, then counts and bets 'orthodox' at times.
well, say while flat betting one experiences a lot of luck and wins a load of money, way above the expectation of a perfect counter. then he turns around and puts in some 'serious orthodox' counting sessions.
question being, is the guy now playing as a serious counter going to 'pay' for his flat betting luck sessions by losing a load of money to get him statistically where he should be?

No, he shouldn't worry about his previous lucky sessions. He should expect to earn his EV, or at least somewhere within a few SDs, on each session. The past results have no bearing on his future results.

-Sonny-

 

[kewljason]

Ah, the Gambler's Fallacy rears it's head again. The player is not "due" for anything other than his EV (plus or minus the variance). The results can "even out" even if you never have another losing session for the rest of your life.

No, he shouldn't worry about his previous lucky sessions. He should expect to earn his EV, or at least somewhere within a few SDs, on each session. The past results have no bearing on his future results.

-Sonny-

I learned this fact during my second year of counting. I started the year on a great run. Up 4 grand after 20 some hours, playing with a $10 unit. I forget the detail, but something like times above EV. I figure I was due for a losing period, so why not cut my unit to $5 and only lose half as much. ($5 were a bit more pleniful than they are now, but still hard to come by) Anyway that losing streak never came. Over the next several hundred hours, I won at a much smaller rate but still won. I was ahead like $5400 after 300 hours of so, and my EV was pretty close to were it should have been. What I managed to do was cut my win total in half, winning $1400 rather than $2800 over that time. It was a good lesson for me though. Always believe in the math.

 

[sagefr0g]

another question on this stuff and flat bet basic strategy

ok, another question regarding playing just basic strategy, maybe using indices, but just flat betting against some game conditions.

mostly except maybe some exotic single deck game it's gonna be a negative ev game.
i notice like in cvcx, if you flat bet just playing basic strategy with maybe using i18, that there is not a number for hands to N0 posted.
why would that be?
am i understanding correctly, that number of hands to reach N0 is the number of hands one would play to reach the point where one's expectation and one standard deviation become essentially equivalent?
like flat betting and playing basic strategy, one would expect over the 'long haul' to lose some amount per hour, or per say 100 hands. but what one see's is this series of up and down sessions, sort of thing.
is there no number of hands for which one gets to the point where one sees that up and down stuff, reach a 'leveling off' sort of thing where over all it's to the point where your losing what you expect to lose, sort of thing?
maybe i'm just confused on what this idea of N0 means?:whip:

 

[Kasi]

i notice like in cvcx, if you flat bet just playing basic strategy with maybe using i18, that there is not a number for hands to N0 posted.why would that be?
am i understanding correctly, that number of hands to reach N0 is the number of hands one would play to reach the point where one's expectation and one standard deviation become essentially equivalent?

You are undersatnding correctly, trust me.

CVCX won't post N0 in -EV games becasue the asumption behind N0 is that, when a card-counter achieves it, it becomes more reasonable to assume such results are due to the result of "skill" or the successful application of some AP system. Since his EV is positive per hand, the point at which N0 occurs would mean, if he were exactly even in $'s, he would be at the low point of 1 SD but still hasn't lost money. He has suffered "bad luck" to the tune of 1 SD but he is still even. It gives asurance to the player he actually may be properly applying his "system" for that game ie counting and betting and using indexes properly. It's completely artificial but maybe something nice to know nonetheless.

Like if his EV is 3 or more SD's below at N0 point, maybe the odds increase a little he might not not have the "right stuff" and be doing something "wrong" somewhere along the line. Of course, the possibility still exists, he is a perfectly applying his system and is just one unlucky sunuvabitch.

A flat-betting BS guy in a -EV game also has a point at which "N0" is achieved except it doesn't really imply a damn thing since any results he may achieve or not are strictly the result of "luck". I suppose it could be used to deduce how well that flat-betting guy is applying the proper BS to that game or maybe even if the casino was "cheating" or not.

For a flat-betting BS guy, the point at which N0 occurs would mean the high point of 1SD means he is still even in $'s since his EV at that point is always negative.

A BS internet guy like me only always "knew", more accurately made best "guess", how many SD's I was ahead or behind had I flat-bet every hand, even though I never actually did flat-bet every hand. I never had heard of the N0 concept, and, if I had, why would I have cared?

I think my most extreme FB results were being down 3.7 SD's or so at Intercasino after 7 or 8000 hands. That's a pretty bad run of 8000 hands lol. I think after 40 or 50 thousands hands, EV, negative as it obviously was, was at least back to EV.

At that point, I'd say my results were within N0 and it was likely I probably had been applying proper BS for the most part and the casino was not actually dealing crooked cards to me lol.

Take one of my "play-all" sheets and make all units 1. EV will become negative but there will still be an N0 number of hands - the point where EV, negative as it is in units or $'s, eguals the amount of 1SD.

 

[sagefr0g]

....
Take one of my "play-all" sheets and make all units 1. EV will become negative but there will still be an N0 number of hands - the point where EV, negative as it is in units or $'s, eguals the amount of 1SD.

ahh, ok, i should of thought of that. thank you and thanks for the other explanations as well.

just, i'm curious as to an opinion on the matter of flat betting basic strategy stuff and that other question i asked in the link:
http://www.blackjackinfo.com/bb/showpost.php?p=141020&postcount=21

i mean, i dunno, flat betting basic strategy, you know it's a 100% ROR proposition.
thing is about that, i don't understand what that means as far as when it is that it'll happen that you'll face ruin.
i think i know that you can know what negative expected value you'll be plugging along at. so if you know you have X amount of roll, then because you know the ev you can know how many hands you can expect to play before ruin. then you can know the standard deviation involved as well so you have an idea of how your roll can go up and down, just maybe not when or to what degree that fluctuation will occur. like it could be 1sd, 2sd, 3sd, ect.
thing is, it would seem as if how that standard deviation stuff 'came down' that it would affect the actual number of hands it would take you to reach ruin. like maybe you just have a lot of good luck early on, then it seems it would be likely that it would take longer before your were ruined and vice a versa for if you had a lot of bad luck early on, sort of thing.
thing is i guess one could have all kinds of degree's of luck in the middle and later stages of one's progress towards ruin, lol.
so i guess how it's supposed to be considered as far as the standard deviation stuff is, that your supposed to look at it as being sort of equally likely over time on either side of what the ev is. so you just have an idea of a range of ways that you could end up getting ruined.
so i guess the 100% ROR means that sooner or later (according to the percentages of the levels that sigmas occur) about some expected number of hands is when ruin will occur.
but anyway, i wonder about a scenario like follows (that is similar to the question in the link above):

say your flat betting basic strategy and you have a really nice run of luck, make a nice little sum of money for some session.
ok does that mean that some where along the line because of that run of luck that some where along the line in the future, that the laws of probability are going to 'make up' for that run of luck and drag you down? <<== edit: ahh so this is gambler fallacy
i mean i think from the stuff Sonny posted, that really, that's not how it works. i mean, yes, lmao, some where on down the line your gonna get dragged down, but that's just the way it is and it has nothing to do with the fact that previously you'd had a nice run of luck.

so but still i think i understand that the more hands you play flat betting basic strategy, that regardless of what happened previously, your getting closer to ruin because of the fact of getting more hands in. maybe there is a standard deviation as far as how many hands it might take to reach ruin?
theoretically it could take a lot more or less hands to reach ruin than the expected number of hands?

is that pretty much how it is or am i missing something?:whip:

edit: well anyway, looking at your 'play all sheets' for my typical game it looks like 89851 rounds would get me to N0.
there supposedly i'd expect to have lost $1,710.00 .
those number of rounds would be about 898 hours of play.
typical for me would be about 3 hours a session or less.
that would be almost 300 sessions.
i think i've played about 237 sessions in four years, so far and i'm good for six grand won.
so if i just flat bet basic strategy say about three hour sessions at the rate of sessions per year i've been playing, it seems i should be good for about 900 of my typical three hour sessions or about fifteen years of play give or take i dunno how many years.
so maybe if i did that, it might cost me about $400 a year to play,or just under $7 per session.
i'm not sure if i'm figuring all that correctly.
anyway, if i did that would it still be possible to not lose all that money and still play that way? i mean with luck could it be less money lost, or with bad luck maybe more money lost?
am i even figuring this stuff anywhere close to right?
lucky if i should live so long, lol.

 

[Kasi]

i mean, i dunno, flat betting basic strategy, you know it's a 100% ROR proposition.
thing is about that, i don't understand what that means as far as when it is that it'll happen that you'll face ruin.
i think i know that you can know what negative expected value you'll be plugging along at. so if you know you have X amount of roll, then because you know the ev you can know how many hands you can expect to play before ruin. then you can know the standard deviation involved as well so you have an idea of how your roll can go up and down, just maybe not when or to what degree that fluctuation will occur. like it could be 1sd, 2sd, 3sd, ect.
thing is, it would seem as if how that standard deviation stuff 'came down' that it would affect the actual number of hands it would take you to reach ruin. like maybe you just have a lot of good luck early on, then it seems it would be likely that it would take longer before your were ruined and vice a versa for if you had a lot of bad luck early on, sort of thing.
thing is i guess one could have all kinds of degree's of luck in the middle and later stages of one's progress towards ruin, lol.
so i guess how it's supposed to be considered as far as the standard deviation stuff is, that your supposed to look at it as being sort of equally likely over time on either side of what the ev is. so you just have an idea of a range of ways that you could end up getting ruined.
so i guess the 100% ROR means that sooner or later (according to the percentages of the levels that sigmas occur) about some expected number of hands is when ruin will occur.

Well, flat-betting or not, lifetime ROR will still be 100% lol. The thing to keep in mind is SD/hd is fixed. I prefer to fix it in units but so what. SD can only be measured over a fixed number of hands. If it's, say, 1.15 unit per hand while flat-betting, then it's 11.5 units over 100 hands. 3SD would be + or - 34.5 units over the same 100 hands. If you have 15 unit roll, you just have less of a chance of surviving the next 100 hands should Lady Luck turn against you in those 100 hands. If you have a 110.5 unit roll it would take a -10SD event for you to lose all in the next 100 hands. One can always measure the liklihood of losing the remaining units in roll over x number of future hands.

While one can't know when or to what degree fluctuations may occur, one does know the liklihood of such events occuring in the next x hands. It will always be, for example, 11.5 units over the next 100 hands.

but anyway, i wonder about a scenario like follows (that is similar to the question in the link above):

say your flat betting basic strategy and you have a really nice run of luck, make a nice little sum of money for some session.
ok does that mean that some where along the line because of that run of luck that some where along the line in the future, that the laws of probability are going to 'make up' for that run of luck and drag you down? <<== edit: ahh so this is gambler fallacy
i mean i think from the stuff Sonny posted, that really, that's not how it works. i mean, yes, lmao, some where on down the line your gonna get dragged down, but that's just the way it is and it has nothing to do with the fact that previously you'd had a nice run of luck.
so but still i think i understand that the more hands you play flat betting basic strategy, that regardless of what happened previously, your getting closer to ruin because of the fact of getting more hands in. maybe there is a standard deviation as far as how many hands it might take to reach ruin?
theoretically it could take a lot more or less hands to reach ruin than the expected number of hands?
is that pretty much how it is or am i missing something?:whip:

I don't know - pretty much makes sense to me lol. Like I said, there always is a SD that will create ruin over the next x fixed number of hands. It's the lilkihood of such an event occuring that tells one the lilklihood of busting - if it would take a 7 SD event to lose all your roll over the next 25000 hands, then you are probably pretty safe for the next 25000 hands. If it would take only a -0.5SD event to lose all over the next 25000 hands, then you are not so safe lol. The only think you know is how many units constitute 1SD over the next 25000 hands and how often 1SD events occur which is fixed. In this case 1.15 units times sqrt of 25000=182 units. So, if you still have 4 times 182 units in your roll, you have a good chance of surviving the next 25000 hands since -4SD events don't occur that often.[/QUOTE]

edit: well anyway, looking at your 'play all sheets' for my typical game it looks like 89851 rounds would get me to N0.
there supposedly i'd expect to have lost $1,710.00 .
those number of rounds would be about 898 hours of play.
typical for me would be about 3 hours a session or less.
that would be almost 300 sessions.
i think i've played about 237 sessions in four years, so far and i'm good for six grand won.
so if i just flat bet basic strategy say about three hour sessions at the rate of sessions per year i've been playing, it seems i should be good for about 900 of my typical three hour sessions or about fifteen years of play give or take i dunno how many years.
so maybe if i did that, it might cost me about $400 a year to play,or just under $7 per session.
i'm not sure if i'm figuring all that correctly.
anyway, if i did that would it still be possible to not lose all that money and still play that way? i mean with luck could it be less money lost, or with bad luck maybe more money lost?
am i even figuring this stuff anywhere close to right?
lucky if i should live so long, lol.

wELL, i'M JUST GONNA ASSUME YOU HAVE A $5 UNIT AND sd IS 1.15 UNITS/RD. Whoops, sorry for the caps. Maybe a -EV of 0.38% or so.

Then 1SD in 89851 rds is about 342 units.

If you have picked up an additional 1200 units to this point, that's cool.

ROR is a function of unit roll, EV and SD. I think.

Of course it's quite possible with a large enough unit roll you may have a 0%ROR from this point forward given a limited lifetime.

Keep in mind you may be playing more or fewer than 300 hands in your 3 hour sessions. To me it's about number of hands played - you still have to play 89851 hands to achieve N0 whether it takes 898 hours or 600 hours or 1200 hours.

Also, I'm ignoring the chances of "finishing" x units behind vs chances of losing all before that but, had you had more roll, you may have been able to recover.

Maybe try using some of QFIT's calculators to determine chances of finishing wherever in so long to see how safe your current roll maybe for the next 50,000 hands or whatever.

But, hey, if it's -0.0038 EV, you lose 1 unit in 263 hands. You'd lose a 1000 unit roll in 263000 hands, simplistically speaking.

Here's your homework assignment - how many SD's ahead are, and what are the chances of that happening, you by winning $6K in 237 "sessions" in this game?

Assuming flat-betting I guess since that's what you seem to imply you have been doing.

But I already know, ok I think I know lol, you must have won a few hands at more than the "flat-bet" unit.

In response to your link, you know how they have that blindfold over some statue in front of court houses to imply "Justice" is blind and equal to all. Well, Lady Luck wears that same blindfold and I guarantee you she does not discriminate based on age, sex, sexual orientation, religion, ethnicity, wealth or intelligence or whether you are AP or not. Not so sure about the "Justice" statue.

 

[sagefr0g]

home work assignment, lol

wELL, i'M JUST GONNA ASSUME YOU HAVE A $5 UNIT AND sd IS 1.15 UNITS/RD. Whoops, sorry for the caps. Maybe a -EV of 0.38% or so.
.....

yup you got it. well close enough -0.37ev and sd 1.14unit/rnd or so,.... i think the sim was using I18.

Here's your homework assignment - how many SD's ahead are, and what are the chances of that happening, you by winning $6K in 237 "sessions" in this game?

Assuming flat-betting ...

yer beginin to remind me of a certain third grade school marme.:flame::whip::joker:

going by circa 3 hours per session, 100 hands per hour, flat betting, .... maybe not really how it was but just say if it was that way for 237 sessions.

237session x 3hr/session x 100hands/hr = 71,100hands

might as well say four standard deviations to the good from your play all sheet for 71,100 hands, cause one standard deviation to the good would be $1,521. where four sd would be 4 x $1,521 = $6,084, close enough.
so the chances of that happening are not so likely since it's a 4sigma event.
just all i think i know on the chances of it, is like a 1sigma deal to the good would be about 34.1% of the time, 2sigma deal to the good about 13.6% of the time, 3sigma deal to the good about 2.1% of the time and i think 4sigma to the good would be about 0.1% of the time.
would the odds of it happening be 0.001 to 1, like once in a thousand tries for trying 71,000 hands worth of play, sort of thing?
so maybe i was really doing ok when i was counting cards over a good bit of those sessions. must have had some pretty good luck on that too, lol.

edit: probably another way to figure that stuff, but i don't know how....:fish::fish:

 

[Kasi]

yup you got it. well close enough -0.37ev and sd 1.14unit/rnd or so,.... i think the sim was using I18.

yer beginin to remind me of a certain third grade school marme.:flame::whip::joker:

going by circa 3 hours per session, 100 hands per hour, flat betting, .... maybe not really how it was but just say if it was that way for 237 sessions.

237session x 3hr/session x 100hands/hr = 71,100hands

might as well say four standard deviations to the good from your play all sheet for 71,100 hands, cause one standard deviation to the good would be $1,521. where four sd would be 4 x $1,521 = $6,084, close enough.
so the chances of that happening are not so likely since it's a 4sigma event.
just all i think i know on the chances of it, is like a 1sigma deal to the good would be about 34.1% of the time, 2sigma deal to the good about 13.6% of the time, 3sigma deal to the good about 2.1% of the time and i think 4sigma to the good would be about 0.1% of the time.
would the odds of it happening be 0.001 to 1, like once in a thousand tries for trying 71,000 hands worth of play, sort of thing?
so maybe i was really doing ok when i was counting cards over a good bit of those sessions. must have had some pretty good luck on that too, lol.
edit: probably another way to figure that stuff, but i don't know how....:fish::fish:

Well, once you told me N0 was $1710 and 89K~ hands it wasn't hard to figure out $unit and EV lol.

I'll match my 7th grade Latin teacher against your 3rd grade marme anytime lol. This woman, while about 300 years old and 80 pounds, remains the most frightening thing in my childhood psyche to this day. I did learn Latin though.

I only gave you the homework assignment becasue, when I asked it, I was coming up with astronomically impossible probabilities lol.

So, I'm glad I asked becasue you are right - today it seems more like the 4 SD's you say than the 10SD's I was coming with last nite somehow lol.

However, you only get partial credit for, while answering correctly your results were 4 SD to the good, you may have messed up a little on the likelihood of this happening.

1sd means one's results will fall within 1 SD 68.27% of the time. Which in turns means 31.73% of the time one's results will not be within 1SD. Which in turns means half of that 31.73% they will be to the left, bad, and half to the right, good, of the curve.
Similarly,
2SD=95.45%.
3Sd=99.73%.
4SD=99.99366%.
of the time one's results will fall within those ranges.

So for 4SD, half the time of (1-.9999366) one's results be to the good of (1-9999366) and the other half to the bad.

In other words, the chances of you doing better than what you did are 1/0.00003168= 1 in 31560.

Another way is just use the normsdist function in Excel but keep in mind it will mean "chances of doing at least this good" since it's "cumulative". That's the way I understand it anyway.

So "=normsdist(1)" will equal 0.8414 but "=normsdist(-1) will equal 0.1586. So, the former means "84% of the time my results will be +1SD or lower" while the latter means "15.86% of the time my results will be this bad or worse".

Maybe you could use QFIT's widget area in cvcx to plugin 71100 hands in or 711 hours with a goal of 1200 units won or something and see what it says. Maybe as a "trip" session lasting 711 hours or something .While maybe giving yourself a generous starting roll. I've never completely understood alot of those numbers in those sections lol.

After winning these 1200 units, how many units do you consider are in your current roll from this point forward? Now that you know SD in units per hand, and EV in units per hand, you can probably always use one of his calculators to determine chances of finishing wherever over the next number of hands.

Go figure, here in voodoo, using both a sim and math to help a voodoo guy, I consider a FB BS guy a voodoo guy using a betting system of 1 unit every hand, answer the question of how long can I expect my sytem to last and how likely have my results been so far.

Now go home and show your Mom and Dad that B+ you got on your BJ test :grin:

 

[sagefr0g]

.....

However, you only get partial credit for, while answering correctly your results were 4 SD to the good, you may have messed up a little on the likelihood of this happening.

1sd means one's results will fall within 1 SD 68.27% of the time. Which in turns means 31.73% of the time one's results will not be within 1SD. Which in turns means half of that 31.73% they will be to the left, bad, and half to the right, good, of the curve.
Similarly,
2SD=95.45%.
3Sd=99.73%.
4SD=99.99366%.
of the time one's results will fall within those ranges.

So for 4SD, half the time of (1-.9999366) one's results be to the good of (1-9999366) and the other half to the bad.

In other words, the chances of you doing better than what you did are 1/0.00003168= 1 in 31560.

.....
Now go home and show your Mom and Dad that B+ you got on your BJ test :grin:

lmao, way worse than my third grade teacher....:joker::whip:
since i know the dog ate my homework routine a'int gonna cut it, i made up the excuse below hoping for extra credit, lol.

but ok i see where i messed up, cause i just used the graph below off of google images to find the figure for 4/2sigma.
thing is i guess that last number 0.1 (the number i used) on the good side is for 4/2sigma + 5/2sigma + 6/2sigma + .... ===>>> infinitysigma/2 maybe.

a B+? :2nd::celebrate
my folks a'int gonna believe this.

thank you for that other stuff and i'll read up on it.

 

[Kasi]

but ok i see where i messed up, cause i just used the graph below off of google images to find the figure for 4/2sigma.
thing is i guess that last number 0.1 (the number i used) on the good side is for 4/2sigma + 5/2sigma + 6/2sigma + .... ===>>> infinitysigma/2 maybe.
a B+? :2nd::celebrate
my folks a'int gonna believe this.

The only reason you got a B+ was because it seemed as if you may have pursued the question with some mis-guided conclusions.

Effort counts in my class, however misguided it may be :grin:

Can't see it now, but it seems like your graph indicated a 68.2% under the curve. It seemed like it indicated a a 27.2% more for 2SD under the curve for a total of 95.4% under the curve.

Whatever it said, it was fine compared to what I said.

Does it look like to you 0.1% is the same length outside 3SD as 2SD was from 1SD to imply it would actually be at or imply 4SD's?

But when effort is combined with such obvious crazy conclusions, when it can clearly be seen the 0.1% is nowhere near enough far away to conclude it represented a 4SD point, when it is so obvious your source could not mean that, I have to revise my grade to a B- for not pursuing the obvious. :grin:

If you want to get your grade back up to a B+, maybe even higher, before I tell your mom and dad, you tell me how many -SD's that 0.1% represents.

Hint, use the Excel function instead of these goofy google graphs. Which are actually fine as far as they go lol.

Dammit, now I'm gonna have to figure it out lol.

 

[sagefr0g]

The only reason you got a B+ was because it seemed as if you may have pursued the question with some mis-guided conclusions.

Effort counts in my class, however misguided it may be :grin:

Can't see it now, but it seems like your graph indicated a 68.2% under the curve. It seemed like it indicated a a 27.2% more for 2SD under the curve for a total of 95.4% under the curve.

Whatever it said, it was fine compared to what I said.

Does it look like to you 0.1% is the same length outside 3SD as 2SD was from 1SD to imply it would actually be at or imply 4SD's?

no, not when i look at it, lol.
my only excuse is, i said it was almost, lmao, not exact.
heck, something that remote a possibility as 4SD's to the good didn't seem like such a big deal.
am i gonna flunk this class? lmao, i should just stick with the B+ :angel::whip:

But when effort is combined with such obvious crazy conclusions, when it can clearly be seen the 0.1% is nowhere near enough far away to conclude it represented a 4SD point, when it is so obvious your source could not mean that, I have to revise my grade to a B- for not pursuing the obvious. :grin:

If you want to get your grade back up to a B+, maybe even higher, before I tell your mom and dad, you tell me how many -SD's that 0.1% represents.

Hint, use the Excel function instead of these goofy google graphs. Which are actually fine as far as they go lol.

Dammit, now I'm gonna have to figure it out lol.

what? that last number 0.1 (the number i used) on the good side is for 4/2sigma + 5/2sigma + 6/2sigma + .... ===>>> infinitysigma/2 isn't the correct answer for how many -SD's that 0.1 represents?
hmm, i'm lost.
ok i'll fool around with Excel functions, not sure what function to even try.:fish::fish:

edit: ahh ok, i think i see, that 0.1% on the graph is only represented by a small part of a sigma, not the whole infinite but ever decreasing area under the curve past the 3sigma point.
so that 0.1% would be at about the 3.25sigma point, judging from the graph.
but i'm still trying to figure out what Excel function to use.:whip:

 

[Kasi]

am i gonna flunk this class? lmao, i should just stick with the B+ :angel::whip:..

Heck no lol.

Aside from the fact I also learn from you and all your questions, I, as a teacher, am not allowed to flunk the "slower" children of America in the current political climate :grin: It makes it appear our billions in education $'s may be wasted lmao.

I'll keep trying as long as you keep trying until one of us understands. And, it may very well be me learning something before it is you.

Type "=NORMSDIST(1)" into a cell in Excel. The result displayed should be "0.84134474."

Type "=NORMSDIST(-1)" into another cell in Excel. The result displayed should be "0.15865526"

Note how the 2 results sum to "1" and the first result less the 2nd resut ="0.68268948", the area under the curve that represents 1SD whether + or - from the mean (think "EV"). What this means is 68.26...% of the time, one's results will fall within 1SD.

It also means, when it says "0.84134474", that 84.13...% of the time one's results will be that good or worse. In effect, it includes all areas to the left of the curve including any results outside -1SD. It's "cumulative" with respect to what's "left" of the curve.

When one uses the "=NORMSDIST" function in Excel it means "my results will be at least this good or better x% of the time" and includes all areas to the left of the curve when the the "x" in "=normsdist(x)" is positive.

When the "x" in "=normsdist(x)" is negative it means, "my results will be at least this bad or worse" x% of the time".

So, type in "=NORMSDIST(-4)" into a cell. The result will be "0.0000316860".
This means both "my results will be at least this bad or worse 0.0031686% of the time" or "my results will be better than this "1- 0.0000316860" (0.999968314) of the time.

It also means the area under a 4SD curve is 0.999936628 and all events will fall outside of this frequency 0.0000633721 of the time, half of that to the right of the curve (0.0000316860) of the time and half (0.0000316860) to the left of a 4SD curve.

As "unlikely" as 4SD is, it's only a 1 in 31560 to do better than that or worse than that and stuff happens. Get to 5SD and it's 1 in 3+MM. Get to 6SD and it's over a billion to 1. Get to 8 and it's 1500 trillion to 1 and you either suck or are being cheated.

If you get some weird scientific notation when you use this, just format the cell to a "number" with 10 decimals or so.

That's the point of measuring one's results, one's "luck" if you will, and how far they may be from expected. One can put a number on how good or bad one's "luck" has been. Well, one can if one has run a sim so one knows one's EV and SD/hd.

Like in Card Counting forum, someone sometimes posts some bad results and everyone says "a little bad luck happens. Don't worry about it, it will come around" even if the results are pretty much unbelievably bad. Likewise, if one posts a big win, it's "expected" even if the win is unbelievably good.

Of course, it must all be in the same game playing the same way, ideally, to have a point to measure the results in the first place.

One shouldn't count those "steaming" $1000 hands when behind and, when won, think "now I'm at EV just like I am supposed to be" lol.

So you were 1 in 31000 "lucky" over those 71100 hands. You could play 71,100 hands 30099 more times and not expect to do better than that. Maybe that's what that means? Assuming of course you were actually playing pretty much the same game pretty much the same way for those 71,100 hands like a Weekend Warrior should lol.

How lucky would you be if it happened you were playing 80 hands an hour instead of 100/hds/hr? Or even crazier 60 hands an hour? Now your results are 1 in a couple hundred thousand "lucky" or 1 in over 5MM "lucky". That's another homework assignment - to verify what I did was OK lol.

Which is why I think one should write down number of rounds played (best guess), even a range of number of hands so later you could develop a high-low estimate, after each "session" rather than only know, much later, how long you may have sat there in hours.

Probably even more important to do this when back-counting when a sim's "hourly" rate assumes you see 100 hds/hr but you only play 13% of them. The good news is, it shouldn't be that hard to know how many physical rounds you may have played in an evening since you have played so few. So don't go back to your room and just write down "3 hours", write down "x number of rounds played" and the 3 hours.

The beauty of this "hands played" way it also allows one to measure one's "luck" over a variety of different games with different pens, decks, spreads, etc if one only knows the EV and SD for each of those games and spreads per round played individually. In effect, a way to measure one's lifetime "luck" over many different games and spreads etc since both EV and variance are additive per round played.

I digress. If you use the "=NORMSDIST(x)" function in Excel the result always includes ALL areas to the "left" of that to infinity.

 

[sagefr0g]

Heck no lol.

Aside from the fact I also learn from you and all your questions, I, as a teacher, am not allowed to flunk the "slower" children of America in the current political climate :grin: It makes it appear our billions in education $'s may be wasted lmao.

I'll keep trying as long as you keep trying until one of us understands. And, it may very well be me learning something before it is you.

Type "=NORMSDIST(1)" into a cell in Excel. The result displayed should be "0.84134474."

Type "=NORMSDIST(-1)" into another cell in Excel. The result displayed should be "0.15865526"

Note how the 2 results sum to "1" and the first result less the 2nd resut ="0.68268948", the area under the curve that represents 1SD whether + or - from the mean (think "EV"). What this means is 68.26...% of the time, one's results will fall within 1SD.

It also means, when it says "0.84134474", that 84.13...% of the time one's results will be that good or worse. In effect, it includes all areas to the left of the curve including any results outside -1SD. It's "cumulative" with respect to what's "left" of the curve.

When one uses the "=NORMSDIST" function in Excel it means "my results will be at least this good or better x% of the time" and includes all areas to the left of the curve when the the "x" in "=normsdist(x)" is positive.

When the "x" in "=normsdist(x)" is negative it means, "my results will be at least this bad or worse" x% of the time".

So, type in "=NORMSDIST(-4)" into a cell. The result will be "0.0000316860".
This means both "my results will be at least this bad or worse 0.0031686% of the time" or "my results will be better than this "1- 0.0000316860" (0.999968314) of the time.

It also means the area under a 4SD curve is 0.999936628 and all events will fall outside of this frequency 0.0000633721 of the time, half of that to the right of the curve (0.0000316860) of the time and half (0.0000316860) to the left of a 4SD curve.

As "unlikely" as 4SD is, it's only a 1 in 31560 to do better than that or worse than that and stuff happens. Get to 5SD and it's 1 in 3+MM. Get to 6SD and it's over a billion to 1. Get to 8 and it's 1500 trillion to 1 and you either suck or are being cheated.

If you get some weird scientific notation when you use this, just format the cell to a "number" with 10 decimals or so.

That's the point of measuring one's results, one's "luck" if you will, and how far they may be from expected. One can put a number on how good or bad one's "luck" has been. Well, one can if one has run a sim so one knows one's EV and SD/hd.

Like in Card Counting forum, someone sometimes posts some bad results and everyone says "a little bad luck happens. Don't worry about it, it will come around" even if the results are pretty much unbelievably bad. Likewise, if one posts a big win, it's "expected" even if the win is unbelievably good.

Of course, it must all be in the same game playing the same way, ideally, to have a point to measure the results in the first place.

One shouldn't count those "steaming" $1000 hands when behind and, when won, think "now I'm at EV just like I am supposed to be" lol.

So you were 1 in 31000 "lucky" over those 71100 hands. You could play 71,100 hands 30099 more times and not expect to do better than that. Maybe that's what that means? Assuming of course you were actually playing pretty much the same game pretty much the same way for those 71,100 hands like a Weekend Warrior should lol.

How lucky would you be if it happened you were playing 80 hands an hour instead of 100/hds/hr? Or even crazier 60 hands an hour? Now your results are 1 in a couple hundred thousand "lucky" or 1 in over 5MM "lucky". That's another homework assignment - to verify what I did was OK lol.

Which is why I think one should write down number of rounds played (best guess), even a range of number of hands so later you could develop a high-low estimate, after each "session" rather than only know, much later, how long you may have sat there in hours.

Probably even more important to do this when back-counting when a sim's "hourly" rate assumes you see 100 hds/hr but you only play 13% of them. The good news is, it shouldn't be that hard to know how many physical rounds you may have played in an evening since you have played so few. So don't go back to your room and just write down "3 hours", write down "x number of rounds played" and the 3 hours.

The beauty of this "hands played" way it also allows one to measure one's "luck" over a variety of different games with different pens, decks, spreads, etc if one only knows the EV and SD for each of those games and spreads per round played individually. In effect, a way to measure one's lifetime "luck" over many different games and spreads etc since both EV and variance are additive per round played.

I digress. If you use the "=NORMSDIST(x)" function in Excel the result always includes ALL areas to the "left" of that to infinity.

yowser!

lol, i think that's enough for me to digest for now.

like i found some reading to do on this stuff.

http://support.microsoft.com/kb/827369

http://cameron.econ.ucdavis.edu/excel/ex23normaltprobabilities.html

http://www.exceluser.com/explore/statsnormal.htm

http://www.mathsisfun.com/data/standard-deviation.html

something puzzles me about that bell curve.
like, i think you can relate the area under the curve to probability, where the probability is p(e) = 1 sort of thing.
so i think that means that the whole area under the curve would be equal to one.
but, i think that area under the curve is really infinite.
so how could it be equal to one?:whip:
maybe it's negative infinity + positive infinity + 1 = 1 ?
or how is the area under the curve related to probability?
is maybe one half of the curve -1 and the other +1, far as probability goes?
then it would be neg infinity + (-1) + 1 + pos infinity = 0
darned i'm confused on how to conceptualize that.
i don't even want to ask how the mean figures into that or ev, lol, that's would be at the center top of the curve wouldn't it?

then ok, i'm wondering about the chart in qfit's cvcx sim.
like below where the column shows standard deviation against true counts.
just i'm wondering if those numbers are really variance or are they actual standard deviation as labeled?
i mean i guess i could figure it out from your spread sheet, but i'm young and easily confused, get lost easily, lol.

 

[Kasi]

something puzzles me about that bell curve.
like, i think you can relate the area under the curve to probability, where the probability is p(e) = 1 sort of thing.
so i think that means that the whole area under the curve would be equal to one.
but, i think that area under the curve is really infinite.
so how could it be equal to one?:whip:
maybe it's negative infinity + positive infinity + 1 = 1 ?
or how is the area under the curve related to probability?
is maybe one half of the curve -1 and the other +1, far as probability goes?
then it would be neg infinity + (-1) + 1 + pos infinity = 0
darned i'm confused on how to conceptualize that.
i don't even want to ask how the mean figures into that or ev, lol, that's would be at the center top of the curve wouldn't it?

then ok, i'm wondering about the chart in qfit's cvcx sim.
like below where the column shows standard deviation against true counts.
just i'm wondering if those numbers are really variance or are they actual standard deviation as labeled?
i mean i guess i could figure it out from your spread sheet, but i'm young and easily confused, get lost easily, lol.

Lots of things puzzle me too about the curve too lol.

I'd guess the area under the curve is infinite and a "limit" that approaches "1" but never really gets there no matter how many million 9's come after the decimal. I suppose, in theory, one could be 1MM SD's to the left and still recover given einough and roll and time.

Just as "lifetime" ROR assumes infinite play which also is impossible.

The curve can be "taller and thinner" or "shorter and fatter" and still be bell-shaped. Which I think only means, like you say, the top of the curve is the mean and, since the curve is symettrical on both sides, half the time results fall to the left and half to the right.

I have no idea why 1SD always means 68%ish of the time one's results will fall within that area of the curve. Maybe sort of, maybe not, but who cares. Just accept it. I accept it like I accept BS - could never figure out the right play on my own but I have faith in others lol. It allows me to put a number on the liklihood of my results being so far from expected, either to the good or to the bad, after any length of time, which is all I really want to do anyway.

So I can hold my own private pity, "why me Lord", party :cry: :grin:

Not that you could tell from my sheets, maybe, but, yes, those numbers are actually SD's assuming flat-betting at each TC. The way I understand them. If they are around 1.14 to 1.17 or so, they are SD.

Watch out when spreading because those numbers won't change much - I think, my guess is, QFIT's software calculates co-variance invisibly to the user and does not even change the SD numbers accordingly. To get variance, you have to assume some co-variance in that case.

In that picture I think I got an avg SD/rd of 1.158 units/rd for the flat-bettor. Just square that to get avg variance per round. Or add a variance column and square each SD at each TC if you want.

I guess it might be more accurately stated that your pic might be for a flat-bettor using some number of indexes maybe lol. With a cut-card lol. Maybe with H-Lo lol. Whatever you told the sim to do. It's probably a little unlikely, your EV% would exactly equal a "FB BS off the top of each shoe" EV%.

Say, can u make QFIT's stuff say how likely finishing with your 1200 unit was in those, was assuming those 71000 hands, whatever it was, assuming FB etc in your game?

 

[sagefr0g]

...

So I can hold my own private pity, "why me Lord", party :cry: :grin:

well thank you for your take on that stuff.
i think what you was saying makes more sense than the guess's i was trying to make about that stuff.
but anyway, yeah pity party, lol.
like yeah, me, i'll play basic strategy and indices by the book, then when it comes to betting, well that's another story. go figure, lol.:juggle::whip:
like last night, i played for maybe six hours and actually counted orthodox, thing is just one 'dream' shoe presented. what did i do? lol, only could bring myself to spread 1-4. :cat::whip:
hit a blackjack with my 'big bet' of twenty bucks too, lmao, do you think i wasn't kicking myself.
made $2.50 over my buy in for that circa six hours. tipped that to the dealer, went over to the craps table, (did an imaginary oscar's grind) after losing an imaginary two units and winning an imaginary one unit, (voodoo i know), i laid a nickle chip on the pass line, thrower rolled a seven and i won a nickle. went home after that, lol.

Watch out when spreading because those numbers won't change much - I think, my guess is, QFIT's software calculates co-variance invisibly to the user and does not even change the SD numbers accordingly. To get variance, you have to assume some co-variance in that case.

when you say spreading, do you mean going to two or more hands?
your not talking about playing one hand, spreading your bet ramp, instead of flat betting, right?

I guess it might be more accurately stated that your pic might be for a flat-bettor using some number of indexes maybe lol. With a cut-card lol. Maybe with H-Lo lol. Whatever you told the sim to do. It's probably a little unlikely, your EV% would exactly equal a "FB BS off the top of each shoe" EV%.

that's right the sim set up using I18.

Say, can u make QFIT's stuff say how likely finishing with your 1200 unit was in those, was assuming those 71000 hands, whatever it was, assuming FB etc in your game?

ok, i had to use a different sim than the one i was referring to about my past play and the 71,000 hands hopeful future play. some how when i upgraded cvcx i lost that other sim.
but, like with cvcx i think you have to go by hours for the calculator that is relevant to this question. so i went with 710 hrs at 100 hands/hr sort of thing.
the image below is best i could come up with for how likely finishing with the 1,200 units might be. i think with those calculators the best one to use is the one that gives you a range of possible results sort of thing. that would be the expected result widget, that gives a range of results, by one, two and three standard deviations, i guess that is.
i just made the goal, doubling the bank, lol, which obviously a'int gonna happen.:cry::whip:
so but anyway, yeah in that 71,000 hands of play (maybe 15 years, or what ever) i could be toast if i hit three standard deviations to the bad, i guess.

 

[Kasi]

well thank you for your take on that stuff.
i think what you was saying makes more sense than the guess's i was trying to make about that stuff.
but anyway, yeah pity party, lol.
like yeah, me, i'll play basic strategy and indices by the book, then when it comes to betting, well that's another story. go figure, lol.:juggle::whip:
like last night, i played for maybe six hours and actually counted orthodox, thing is just one 'dream' shoe presented. what did i do? lol, only could bring myself to spread 1-4. :cat::whip:
hit a blackjack with my 'big bet' of twenty bucks too, lmao, do you think i wasn't kicking myself.
made $2.50 over my buy in for that circa six hours. tipped that to the dealer, went over to the craps table, (did an imaginary oscar's grind) after losing an imaginary two units and winning an imaginary one unit, (voodoo i know), i laid a nickle chip on the pass line, thrower rolled a seven and i won a nickle. went home after that, lol.

when you say spreading, do you mean going to two or more hands?
your not talking about playing one hand, spreading your bet ramp, instead of flat betting, right?

that's right the sim set up using I18.

ok, i had to use a different sim than the one i was referring to about my past play and the 71,000 hands hopeful future play. some how when i upgraded cvcx i lost that other sim.
but, like with cvcx i think you have to go by hours for the calculator that is relevant to this question. so i went with 710 hrs at 100 hands/hr sort of thing.
the image below is best i could come up with for how likely finishing with the 1,200 units might be. i think with those calculators the best one to use is the one that gives you a range of possible results sort of thing. that would be the expected result widget, that gives a range of results, by one, two and three standard deviations, i guess that is.
i just made the goal, doubling the bank, lol, which obviously a'int gonna happen.:cry::whip:
so but anyway, yeah in that 71,000 hands of play (maybe 15 years, or what ever) i could be toast if i hit three standard deviations to the bad, i guess.

And thank you Wise One.

I do realize you are "betting-challenged" lol. If you did spread 1-4 in that shoe or whatever, run a separate sim for that shoe with that spread and keep results separate from other shoes lol.

Yes I did mean spreading to more than 1 hand lol.

The closest thing I see is a chance of 0.00% of achieving a $6K win with a $6Kroll. That might likely halve or so your chances of finishing $6K ahead since you don't have enough roll to recover from if you go bust diring that time.

I was hoping to see something like 0.02 or 0.03% lol.

Can you run that agian but with a $12K roll and $6K goal bringing all $12K to your trip?

You don't have to show the picture but I was just wondering if that might change that 0.00% chance.

It's tough dealing with such small numbers lol.

I guess basically it's telling you you're results bordered on "impossible" lmao.

 

[sagefr0g]

And thank you Wise One.

I do realize you are "betting-challenged" lol. If you did spread 1-4 in that shoe or whatever, run a separate sim for that shoe with that spread and keep results separate from other shoes lol.

Yes I did mean spreading to more than 1 hand lol.

The closest thing I see is a chance of 0.00% of achieving a $6K win with a $6Kroll. That might likely halve or so your chances of finishing $6K ahead since you don't have enough roll to recover from if you go bust diring that time.

I was hoping to see something like 0.02 or 0.03% lol.

Can you run that agian but with a $12K roll and $6K goal bringing all $12K to your trip?

You don't have to show the picture but I was just wondering if that might change that 0.00% chance.

It's tough dealing with such small numbers lol.

I guess basically it's telling you you're results bordered on "impossible" lmao.

yeah, lol, betting challenged.
but it works both ways, i was happy as a lark going home with five bucks extra, lmao.

it's still the same for $12K rolls and 6K goal except inf went to 0.25 .
i even tried twenty four grand, and no change for the 710 hours.
probably need to throw some voodoo in there to get some action, lol.
almost 6% chance if the roll is six grand and the goal three (erhh edit: make that two) grand for the 710 hours.

but yeah, me counting most of those prior rounds must have helped impossible as it woulda been flat betting.

 

[Kasi]

yeah, lol, betting challenged.
but it works both ways, i was happy as a lark going home with five bucks extra, lmao.

it's still the same for $12K rolls and 6K goal except inf went to 0.25 .
i even tried twenty four grand, and no change for the 710 hours.
probably need to throw some voodoo in there to get some action, lol.
almost 6% chance if the roll is six grand and the goal three (erhh edit: make that two) grand for the 710 hours.

but yeah, me counting most of those prior rounds must have helped impossible as it woulda been flat betting.

Are you kidding me? Up $5? Just another "W" playing a -EV game lol.

Heck, I'm happy as a lark just breaking even lmao.

Thanks for doing that. I was just hoping lol. Always had trouble exactly understanding some of those % in QFIT's stuff lol.

So what if we got 0.03% and he got 0.00% - he got pretty close anyway :grin: :grin: