Value against the closing line, what figures to aim for?

**antonyp22** · 06-10-14, 08:28 AM

Can anyone help with this? Even with their experience in other sports?

**statnerds** · 06-13-14, 09:14 PM

depends on how efficient you view this ARF Market to be. If it is efficient than long term betting the closer should produce favorable results. i would like to see something in the 65% or better range of beating the closers and would like to see at least 100 games played.

how big of a sample size do you have at this point?

**antonyp22** · 06-14-14, 05:43 AM

Now up to a sample size of 100 plays and have beaten the closer 61% of the time

**statnerds** · 06-18-14, 09:02 PM

Sounds fairly solid then. As far as timing, are you betting openers or closer to game time? And where are you at now that you have got 100 games or so, points wise on the CLV?

**antonyp22** · 06-18-14, 11:41 PM

So now after 108 plays, CLV is at 1.58.

In Aussie Rules Football, games are played on Friday night, Saturday and Sunday (weekend games) and all bets are placed at 10am on the Friday.

**brettd** · 06-19-14, 05:10 AM

You have a 55% model. I have been modelling the AFL for a number of years, so I know about CLV numbers relating to Aussie Rules.

**lamichaeljames** · 06-20-14, 10:21 AM

decent model... I am not familiar with AFL lines.

**donkeyshark** · 09-12-14, 09:09 PM

Originally posted by antonyp22

Now up to a sample size of 100 plays and have beaten the closer 61% of the time

I'm seeing a few opinions in here (not trying to discredit those), but I thought some stats knowledge might help you understand your question a little bit better.

If you have a sample size of 100 Gaussian distributed wagers (which is what this should be), and you'd like to know how significant your results are, you can calculate how many standard deviations from the mean you are by taking your successes less your failures and dividing that by the square root of n.

So, here you have:
61-39 = 22
sqrt(100) = 10

sigma = 22/10 = 2.2

So you have significance of 2.2 standard deviations

This translates to a probability of ~ 98.6% likelihood that you are skilled at beating the closing line and a 1.4% probability that you've just gotten lucky. You can calculate this using z-score tables, or in Excel as follows:
=NORM.DIST(61,50,5,TRUE)
where 50 is the mean expectation if it were truly random, and 5 is the stdev of 100 coin flips (calculated as the square root of 100 divided by 2).

**PassTheDutchie** · 09-16-14, 10:48 PM

'where 50 is the mean expectation if it were truly random'

i don't think this is a correct assumption.
When you compare the line at 10 am Friday vs closing lines for a period of time,
i would expect a percentage to be equal to the closing line.

although I have no idea how to take this equal percentage in to account, I have a feeling that it would not change the conclusion of above post.

**indio** · 09-17-14, 02:24 PM

Originally posted by PassTheDutchie

'where 50 is the mean expectation if it were truly random'

i don't think this is a correct assumption.
When you compare the line at 10 am Friday vs closing lines for a period of time,
i would expect a percentage to be equal to the closing line.

although I have no idea how to take this equal percentage in to account, I have a feeling that it would not change the conclusion of above post.

Learn some math. All he is saying is that there is a 1.4% chance (it's actually 1.76%) of randomly hitting 61%+ of 100 coin flips, which although that doesn't sound like much, is a lot. Now, the chances of hitting 61% or more of 400 coin flips is 0.0006% which is approx. odds of 166,665-1. So all he is saying is before the poster wants to have a statue erected, it would be a good idea to have a larger sample size to get a more honest evaluation of abilities vs market.

**ChuckyTheGoat** · 09-17-14, 06:31 PM

Props to you on your endeavor. Sounds pretty good.

One word of caution, re: AFL. Such a hi-scoring league. +1.50 MIGHT not be that great.

**PassTheDutchie** · 09-17-14, 08:58 PM

Learn some math?

My point was that beating the closing line is not comparable with a coin flip and that time of wagering versus game time is probably a factor as well.

**ChuckyTheGoat** · 09-17-14, 09:26 PM

Antony, I'm optimistic for u. Your work should be rewarded. My observation is that AFL lines are tied closely to injury news. Maybe you're ahead of the curve on that?