Arbusers

I am building an arbitrage bot between soft bookmakers ( yes ...i know there should be a sharp one involved in arbitrage but for the markets i am targeting no bookmaker is really 'sharp').

I am only planning to bet the 4%+ arbitrages. However I am expecting around 30% of my arbitrages to fail ( bet fails for one or both legs of the arbitrage ). For these failed arbitrages I will be in one of the following 2 cases:

1) both bets on arbitrage legs failing . This means my bot not losing any money.
2) i manage to place a bet on only one leg of the arbitrage:
  a) the valid bet is on the leg offering oversized odds compare to real probability => VALUEBET?! ..this means I am supposed to win on average more than 4% on this bet right?
  b) the valid bet is on the 'kind of sharp bookmaker' odds => this means I am going to loose the bookmaker juice ...4%-5%-6% on average.

If chance of happening between  a) and b) are 50%/50% does that mean I have a good chance to break even on my failed abitrages?

It depends on the bookies you use. Theoretically if you use sharp vs soft and bot fails to put bet on soft and puts it only on sharp you will lose money in long run. But you dont use sharp so i think most of your not covered bets are value bets. So you should be safe. But why not go totally value so you do not have this handicap ?

Be careful assuming chances are 50%/50%. It seems it could be more likely for your bot to fail on the value leg, since those are where bookmakers will usually try to adjust faster. If one leg has no value at all, there's no reason for the bookie to move odds as quickly.

Wolfie wrote: It depends on the bookies you use. Theoretically if you use sharp vs soft and bot fails to put bet on soft and puts it only on sharp you will lose money in long run. But you dont use sharp so i think most of your not covered bets are value bets. So you should be safe. But why not go totally value so you do not have this handicap ?

I guess going totally value its the way to go. But since I have no real sharp bookmaker I hope that I could still detect the value leg of an arbitrage the following way with a high degree of confidence:
  Once i detect an arbitrage opportunity  I could just bet on the (Bookmaker,Outcome) pair corresponding to the highest percentage increase of odds for the same outcome between Bookmaker1 and Bookmaker2.

barbero wrote: Be careful assuming chances are 50%/50%. It seems it could be more likely for your bot to fail on the value leg, since those are where bookmakers will usually try to adjust faster. If one leg has no value at all, there's no reason for the bookie to move odds as quickly.

You are perfectly right on this . I am still gathering data for doing some statistics on these percentages.

Can you let us know how statistics are looking ?