Byron
Byron
  • Threads: 1
  • Posts: 5
Joined: Jan 18, 2014
January 18th, 2014 at 5:25:28 PM permalink
Can you help me with a BlackJack math question?

I am sure it is an elementary percentage conversion question but I want to make sure my logic and equations are correct.

I believe that I may not be using the statistics in Appendix 4 "Standard Deviation in BlackJack" in wizardofodds.com correctly.

I want to express the number of BlackJack Losses in a row in an easy to understand sentence (stated below).

"You will experience X losses in a row approximately 1 in every Z hands"
X = any number of losses in a row to be computed
Z = the approximate number of hands it takes to incur that loss in a row

Appendix 4 "Standard Deviation in BlackJack" data that I am using:
Probability of a Win 42.42%
Probability of a Loss 49.09%
Probability of a Push 8.48%

I don't want to count pushes so to determine the actual percentage without pushes, I used the below equation:
49.09/(42.42 + 49.09) = .5364
Answer: 53.64%

Now to determine the possibility of x number of losses in a row.
For the example of 10 losses in a row, I used the equation:
.5364^10 = .000250
Answer: .025%

To convert the percentage to odds I used the equation:
.00025/100 = 400,000

Does this mean I can express the result as:
"You will experience 10 losses in a row approximately 1 in every 400,000 hands"?

Is this correct?
Please correct any mistakes I made in my logic and/or formulas.
Thank you,
dwheatley
dwheatley
  • Threads: 25
  • Posts: 1246
Joined: Nov 16, 2009
January 18th, 2014 at 5:31:33 PM permalink
It's 1/0.00025, which should just give you 4000 hands.

And then your final statement is a common but important mistake. Search for 'streak' in this forum, and you will find dozens of people who have asked similar questions, and received detailed responses.
Wisdom is the quality that keeps you out of situations where you would otherwise need it
7craps
7craps
  • Threads: 18
  • Posts: 1977
Joined: Jan 23, 2010
January 18th, 2014 at 6:15:56 PM permalink
Quote: Byron

I don't want to count pushes so to determine the actual percentage without pushes, I used the below equation:
49.09/(42.42 + 49.09) = .5364
Answer: 53.64%

agree. close enough

Quote: Byron

Now to determine the possibility of x number of losses in a row.
For the example of 10 losses in a row, I used the equation:
.5364^10 = .000250
Answer: .025%

wrong
start over
what calculator did you use?

I get about 0.0019719127 in win7 calc
and 1 in 507.12 attempts

Quote: Byron

Does this mean I can express the result as:
"You will experience 10 losses in a row approximately 1 in every 400,000 hands"?

streaks are in attempts
the average length (number of hands or trials) of each attempt for 10 in a row is
2.152778445
(1+p+p^2+p^3+...+p^9)
507.12*2.15 gets the average

1,091.7210
average number of hands (not rounds) to lose 10 in a row
so does (1-(p^run))/(q*(p^run)) where q=1-p

average does not mean 100%, or even 50% - the median is 759
about a 63% chance of at least 1 such run in 1092 hands - not counting pushes
of course it could happen 2 times or more
(standard deviation = 1083.325947)
EventRun Probability1 in
0 runs of length 10 or more0.3676146232.72
at least 1 run of length 10 or more0.6323853767759411.58
at least 2 runs of length 10 or more0.2609227677431593.83
at least 3 runs of length 10 or more0.07688315739874313.01
at least 4 runs of length 10 or more0.01729025648818257.84
at least 5 runs of length 10 or more0.003106594477473321.90
at least 6 runs of length 10 or more0.0004605984023922,171.09

a quick check from a small simulation
shows the values to be quite good
grouped data
items: 100000

minimum value: 10.00
first quartile: 318.00
median: 758.00
third quartile: 1511.00
maximum value: 12344.00

mean value: 1090.07
midrange: 6177.00

range: 12334.00
interquartile range: 1193.00
mean abs deviation: 796.33

sample variance (n): 1168410.46
sample variance (n-1): 1168422.14
sample std dev (n): 1080.93
sample std dev (n-1): 1080.94


use the 2nd calculator here to see the distribution
Streak Calculators
also looks to be very accurate (one can change the value of p in the Excel calculator)

Sally has some data in tables in this thread
on-average-how-many-trials-will-it-take

why do you want this information?
Good Luck
winsome johnny (not Win some johnny)
Byron
Byron
  • Threads: 1
  • Posts: 5
Joined: Jan 18, 2014
January 19th, 2014 at 12:32:09 PM permalink
Quote: dwheatley

It's 1/0.00025, which should just give you 4000 hands.

Byron
Byron
  • Threads: 1
  • Posts: 5
Joined: Jan 18, 2014
January 19th, 2014 at 12:33:33 PM permalink
Thanks for that correction to my equation.
Byron
Byron
  • Threads: 1
  • Posts: 5
Joined: Jan 18, 2014
January 19th, 2014 at 12:42:17 PM permalink
Quote: 7craps

agree. close enough

wrong
start over
what calculator did you use?

I get about 0.0019719127 in win7 calc
and 1 in 507.12 attempts

My reply:
I used MS Excel 2010 to calculate, but sorry had a stupid typo....used .4364 instead of .5364.
You know what they say "bad data in, bad data out"

Also sorry I don't know how to reply and get the Quotes to show up in the blue box and my reply to show up in white.

Byron
Byron
  • Threads: 1
  • Posts: 5
Joined: Jan 18, 2014
January 19th, 2014 at 1:05:54 PM permalink
Your comments:
streaks are in attempts
the average length (number of hands or trials) of each attempt for 10 in a row is
2.152778445
(1+p+p^2+p^3+...+p^9)
507.12*2.15 gets the average

1,091.7210
average number of hands (not rounds) to lose 10 in a row
so does (1-(p^run))/(q*(p^run)) where q=1-p

My question:
Thank you for the above formulas, final questions....
P=probability which is my case is 0.0019719127 , correct?
What would the value of "run" =? Is that my example of 10?
7craps
7craps
  • Threads: 18
  • Posts: 1977
Joined: Jan 23, 2010
January 19th, 2014 at 1:16:33 PM permalink
Just to quote me, have the closing tag [ / q ] before your reply

and everything to me is in blue

Quote: Byron

My question:
Thank you for the above formulas, final questions....
P=probability which is my case is 0.0019719127 , correct?
What would the value of "run" =? Is that my example of 10?

No
P<> 0.0019719127
p= 0.5364 (probability of a loss)

yes, run = 10 = the length of the streak. your example of losing 10 in a row

p^10 = 0.0019719127
or
1/ p^10 = 507.12 = the average number of attempts (not hands)

hope this all helps
winsome johnny (not Win some johnny)
  • Jump to: