bets4value

Other Entries

Monkeys Tossing Coins

Apologies to the well respected Joseph Buchdahl from www.football-data.co.uk for stealing the title for this article. I find it appropriate and creative at the same time so couldn't resist using it. I swear the rest of the article are my very own words... :-) 

It is oh-so-tempting to check a half-decent tipster-platform like blogabet, betadvisor, pyckio, etc. and sign up with a tipster who is being promoted as the next big thing. Even for more experienced punters it is sometimes hard to resist.

Just out of curiosity I fired up my browser and checked the blogabet site which read the following stats for the tipster topping the table: 893 bets, 19% yield, 9524 followers.

Damn it, that’s too good to be true, isn't it? For someone to show a 19% yield over such a ‘large’ sample he must have what it takes, right? Also if he continues that performance the €50 for monthly sub-fees are covered in no time…

But hold on. Are we missing something?

YES, WE ARE!

Let’s do a little experiment and I’ll show you what you we are missing. Specifically I will show you that I’m able to craft a successful tipster out of nothing!

Let’s say we have a pool of 100 tipsters. Each tipster is entitled to 1000 bets on 1000 football matches – one bet for each match only (home win, draw or away win). All matches are real historical matches with pinnacle closing prices. A small sample of the matches could look like this 

 

Screen Shot 2017-11-13 at 20.18.52

 

There is also one secret that we must not tell the audience for now: Every tipster is generating his bets completely randomly. 

After every tipster has placed his bets they are then settled to actual results and profit is recorded. Here is the table of the 10 best performing tipsters. Btw. we have dumped the rest as we are not interested in the losers, are we?

 

Screen Shot 2017-11-13 at 20.37.10

 

Doesn’t that look great? The top-10 tipsters all show a profit! Taking the top tipster with a 11.74% yield over a 1000 bet sample must be a sound investment, right?

What if I’d tell you now that all those tipsters were simply tossing coins as per our assumption. It doesn’t feel like a such great investment anymore, eh?

What the heck is going on here? The best tipster has achieved a yield of 11.74% over 1000 bets by guessing, in fact he was making 1000 bets with negative expectation yet he is still showing a profit. Statistically, that is not at all surprising! Probability dictates, that out of the tens of thousands of people tipping, some at least are going to get it right due to the sheer volumes.

This is what is widely known in the betting industry as Survivorship Bias and traditional tipster platforms all suffer from this phenomenon.

Only successful tipsters shout and boast about their winning selections. The losing tipsters delete their tips and start tipping under a new username or website. By only taking tips from ‘winning’ tipsters you have ignored the fact that they could quite easily have given losing tips. In fact this is exactly what is happening as I showed you with the experiment: every bet from every tipster had negative expectation!

At the point you start following the tips and replicating bets from the most promising tipsters, the previous history has little or no importance and the tipster’s performance will regress to the mean. You will be embracing the same negative expected value going forward and at some point that -ev will inevitably materialize to a loss.

The population of all tipsters do worse than break even, as proven by Joseph Buchdahl in Squares & Sharps, Suckers & Sharks. It’s a pain for me to see how much money is buried in useless tipsters and platforms. Tipping is a classic example of surviorship bias in sports trading. An educated guess would be that 1 tipster out of 10000 indeed genuinely holds an edge over the bookmaker and is able to secure long-term profits. The rest of the bunch are just monkeys tossing coins.

The next time you wonder if something is too good to be true, it probably is.

Posted in Member blogs on November 13 at 08:24 PM

Comments (0)

No login