Identifying Statistical Outliers in my Average Line

[terpkeg]

How do I do it? Is there a standard deviation which my numbers should fall under?

My mathematical and statistical knowledge is fairly weak and I am not sure I have the time and energy for it. But, if it not time consuming or difficult for someone to shed some quick light on this, that would be great.

I have a couple years of tracked lines that I have bet. In 2012 I changed to euro line so excel could give me a mean. Obviously if I bet a future at +4000, it is going to scew my average line and the 40 should be removed. But, what is I make a habit of betting large docs say +350 to +450. At what point am I betting enough of them that they are statistically relevant for my average line. Same thing goes the other way too if I pull the trigger on -400 and up.

[Miz]

I have no idea what you are asking.

[MadTiger]

You have several things you could do. I have seen people GROUP their bets in ranges. +100 to +250 would be a range. -100 to -200 could be another range, and so on. This eliminates your problem, because there would be no outliers; all data points would be valid within their respective group.

[Miz]

oh, i see what you are asking now. First convert from American to decimal. Then you could use median rather than mean. Otherwise after converting all your lines to decimal, you could consider anything 3 standard deviations out as an outlier. A lot of ways to do it. good luck

[terpkeg]

oh, i see what you are asking now. First convert from American to decimal. Then you could use median rather than mean. Otherwise after converting all your lines to decimal, you could consider anything 3 standard deviations out as an outlier. A lot of ways to do it. good luck

Thanks for the input. I converted my excel to decimal as of 1/1/12.

Do you think that using the median accurately reflects my average line v. my win percentage?

So if my median line is 2.2 and I am hitting 48%, can I guaruntee that I am actually profitable? Problem is that every play is not flat bet. So up or down does not necessarily reflect whether I am "winning". If mean line is much higher, it doesnt seem to have any significance with more large futures and long shots than large favorites over the sample size.

[Miz]

I'd really like to help you, but you need to learn how to state your purpose clearly and concisely. For example... how about this:

I am retroactively evaluating the bets I placed over the course of 2012. I am trying to determine the best way to evaluate my performance, and whether or not my bets and success this year are indicative of me being a long term winner. Can anyone help me find the best way to do this?

You have misspelled words and incomplete sentences. It is obvious you took very little time to frame the problem, so why would anyone take a lot of time to respond? Your second reply is much better actually.

I think you need to understand what z-score means. It is a little more complicated when your bets are moneylines that vary, but it can be done. Do a search in here for z-scores. Here is one that I answered for another guy one time... but there is nothing unique to my response, as it has been answered several times before me.

http://forum.sbrforum.com/handicapper-think-tank/366743-sample-sizes.html

Post number 25 should be helpful. Read up on Ganchrow's posts for non point spread stuff. Or Justin's book. Both are good.

[terpkeg]

I'd really like to help you, but you need to learn how to state your purpose clearly and concisely. For example... how about this:

I am retroactively evaluating the bets I placed over the course of 2012. I am trying to determine the best way to evaluate my performance, and whether or not my bets and success this year are indicative of me being a long term winner. Can anyone help me find the best way to do this?

You have misspelled words and incomplete sentences. It is obvious you took very little time to frame the problem, so why would anyone take a lot of time to respond? Your second reply is much better actually.

I think you need to understand what z-score means. It is a little more complicated when your bets are moneylines that vary, but it can be done. Do a search in here for z-scores. Here is one that I answered for another guy one time... but there is nothing unique to my response, as it has been answered several times before me.

http://forum.sbrforum.com/handicapper-think-tank/366743-sample-sizes.html

Post number 25 should be helpful. Read up on Ganchrow's posts for non point spread stuff. Or Justin's book. Both are good.

Yup. Havn-t put the quality time into this that is required. Not sure I ever will, but this looks like it is a good place to start. Thanks.