a bit of mathematics and probability

[zemedelec]

Hi all!

can somebody help me with this?

If I have a tennis match with favourite priced at 1,33 then we can simply calculate that probability of his win is 75%. Is there any way how to calculate probability that he will reach odds of 2,0 ?

Thanx.

[eclipse]

Simply divide 100 by the decimal odds to get the % probability.

Eg. 100/2.0 = 50%

[zemedelec]

THx, but 50% is probability of winning when odds are 2, but what is the probability of move from 1,33 to 2 ? Can this be calculated?

[alrodopial]

Hi all!

can somebody help me with this?

If I have a tennis match with favourite priced at 1,33 then we can simply calculate that probability of his win is 75%. Is there any way how to calculate probability that he will reach odds of 2,0 ?

Thanx.

No, it is not possible.
It depends on current form, conditions etc.
If you want a "close guess" take look at the past performance.

[negapo]

If 1.33 (around 75%) reflects the true probability of the event then it will reach the odds of 2.0 at least 100%-75% = 25%. This is the worst case scenario, where the odds would only reach the 2.0 when the selection would lost (doesn't make much sense, just a reasoning step).

If you closed the trade at 2.0 in that conditions you would have an expected loss of something around 17%.

Of course you would be successful on more than 25% of the times. How much times would you need to be successful is a good indicator.
If you round things for a stake of 100€ you lose 33€ if the selection never reaches 2.0 and you win 33€ if you close it at 2.0. So you need the selection to reach the odds of 2.0 around 50% of the times to reach break even.

And how many times that really happens (the odd moving from 1.33 to 2.00)? As many of the processes that we don't know how are generated we only have a clue of the probabilities by looking at the past and taking it has an indicative of the future probability. So you could download the betfair odds at http://data.betfair.com/ and check the maximum odd with volume above X (a number like 500€) that was taken in-play (there is a flag for bets matched in-play) for the player/team that you are after.

But what does this tell you? That the player X under the physical conditions Y, and many other variables that change, while being the favorite (low odds like 1.33) as tripped and the probabilities have reached something like 2.0.

So, you can only aggregate historical events/probabilities/odds when the process that generates them is stable (at least in part).

[doris_day]

The theoretic probabilities are one thing and the actual matched prices on Betfair quite something else however.

You see this every day in in-play horse race betting. Players on and off-course have pre-set buttons set to prices like 1.10 etc to guarantee they get matched but in the hope that they'll be matched at a much higher price. Most times they do but sometimes the liquidity dries up and they get matched at 1.10 when the 'true' price may be 1.60.

[negapo]

I have the same experience as you doris_day

I see a lot of of ridiculous prices being matched on the exchange (maybe due to errors, liquidity problems, poorly written boots, etc.). I prefer to be on the anti fragility side of the game and hope to gain from this unexpected behaviors.

In the end we are all looking for errors like this or gaps between what a theoretic probability is and the probability we can get our money matched.

You can go either way (or any other way as long as you make money, avoid going broke in the process and can eliminate luck as the reason for your profit).

[gavind]

You can go either way (or any other way as long as you make money, avoid going broke in the process and can eliminate luck as the reason for your profit).

Thanks! I guess it is always preferrable to go slow and steady. It beats everything.