confidence interval and level of certainty

[barry111]

[ QUOTE ]
confidence interval (+/- 5% seems reasonable to me), with a reasonable level of certainty (p value less than .05 seems good)

[/ QUOTE ]

Where do I look on Aleo's spread sheet for the +/- 5%? Does the sheet have p value? I never took stats so this is all foreign to me.
[ QUOTE ]
your confidence intervals between levels can't overlap.

[/ QUOTE ]
Can someone explain what this means?

I don’t post much, so I will use this post to say Thank-You to all the members that add to this fourm. I spend a lot of time read this site everyday and it had helped my game so much.

[/ QUOTE ]

[Irieguy]

I don't have Aleo's spreadsheet in front of me. I'll answer this later today when I can specifically show you where to look.

Irieguy

[SuitedSixes]

On the version I use there's a <CONF> tab, look there.

[barry111]

yes, so on the <conf> tab which value is equal to p? What does Irieguy mean when he said [ QUOTE ]
your confidence intervals between levels can't overlap.


[/ QUOTE ]
Irieguy, I look forword to your response as well
Thanks

[SuitedSixes]

[ QUOTE ]
yes, so on the <conf> tab which value is equal to p? What does Irieguy mean when he said [ QUOTE ]
your confidence intervals between levels can't overlap.


[/ QUOTE ]
Irieguy, I look forword to your response as well
Thanks

[/ QUOTE ]

I don't know, I've never understood a single word that guy has uttered.

[Irieguy]

[ QUOTE ]
[ QUOTE ]
confidence interval (+/- 5% seems reasonable to me), with a reasonable level of certainty (p value less than .05 seems good)

[/ QUOTE ]

Where do I look on Aleo's spread sheet for the +/- 5%? Does the sheet have p value? I never took stats so this is all foreign to me.
[ QUOTE ]
your confidence intervals between levels can't overlap.

[/ QUOTE ]
Can someone explain what this means?

I don’t post much, so I will use this post to say Thank-You to all the members that add to this fourm. I spend a lot of time read this site everyday and it had helped my game so much.

[/ QUOTE ]

[/ QUOTE ]

A "p" value is the percentage chance that your statistical findings are due to randomness alone.

A confidence interval is the range of values you can expect to find 1.00 - p percent of the time.

So when we say "a p value of 0.05," that means there is only a 5% chance that your findings are due to randomness alone. The other 95% of the time, you can expect your findings to fall within a certain range.

On Aleo's spreadsheet, at the bottom of the Poker Summary section at the top left of each stats page, he allows you to plug in how confident you would like to be in your $ per tourney. If you plug in 95%, you will have a p value of 0.05. This will give you a range of $/tourney values.

In order for you to be "sure" that your $/tourney average is higher at one level than another, these ranges cannot include the same numbers.

As the number of SNGs that you play increases, your confidence intervals will shrink. What you will find is that you need an awful lot of SNGs in order for your confidence intervals to NOT overlap between levels. Only then can you tell how your ROIs really compare.

And all of this is only relevant if you believe that you can even apply these statisitcal measures to SNG results. Theoretically, you can't. But there is good emperic evidence that it works.

That should make it as clear as mud.

Irieguy

[barry111]

got it thanks