Historical Betfair BSPs

[LinusP]

From 2012..

[Iron]

Hi Peter

You're right!

Jeff

Jeff
I think you've got that one wrong..
Using your spreadsheet, I make it -£63.94 ROI -12.39%

There is no +1 in the equation as Lian says..Example Take the second row..Ghost Dancer..winning at 1.62 would only give you 58.9p after the 5% commission..not as you have in your spreadsheet

[Iron]

Interesting.

Where did you download that spreadsheet from, out of interest?

Jeff

From 2012..

[NeilSpence]

What happens if you just reverse all bets. Does the curve go the other way??

[Iron]

Logically, if the BSP represents the true price over a very large sample of bets (which it demonstrably does), what will kill you long-term is the commission, so you'll be screwed whether you back or lay!

Jeff

What happens if you just reverse all bets. Does the curve go the other way??

[gutuami]

Anyone know what the average wining odds are for uk horse racing ?

that should help

~3.80 is the average of uk races
~3.90 average for IRE

[LinusP]

I used a firefox addin that allows you to download multiple files at once from a web page then used python to join them together.

Here are the results from laying, as you can see its commission that kills this strategy. If you could trade/bet without paying commission you would be printing money.

[freddy]

Anyone know what the average wining odds are for uk horse racing ?

that should help

~3.80 is the average of uk races
~3.90 average for IRE

Thanks, I thought it was around that.
the 1st spreadsheet had way higher than average winners, the second was about right, but it's a none point now anyway as Jeff got his sums wrong .

By the way Jeff, I once spent days and days going through some Football data and thought i found a great system, only to find out a stupid mistake in my calculations, it easily done

[Iron]

It sure is!

The other challenging thing is that you might think you've found a great system, when in fact you may have hit upon something caused by randomness or a temporary feature of the markets.

Take the graph that Linus posted: viewtopic.php?f=2&t=9801&start=10#p81555.

You might have examined lots of data between about 84,000 and 94,000 or 180,000 and 188,000, and thought you'd come up with a cracking laying system, only to have your dreams destroyed once mean reversion kicked in.

Jeff

By the way Jeff, I once spent days and days going through some Football data and thought i found a great system, only to find out a stupid mistake in my calculations, it easily done

[LinusP]

Just realised my calcs are slightly wrong, need to calculate the profit for each race before applying the commission. Can be done using a sumif I believe.

[Iron]

That makes me feel a bit better about my own miscalculation!

Jeff

Just realised my calcs are slightly wrong, need to calculate the profit for each race before applying the commission. Can be done using a sumif I believe.