Cross Market Arbs

Yeah, pure arbitrage is tricky though, as it tends to be well botted.

My initial strategies on Betfair (football) were purely technical, and purely based on the odds in relation to each other. I wasn't watching the matches I was betting on.

I'm open to correction, but basically, there are no cross market arbs anymore.

Cross market arbs (of the simple risk free variety) have been eliminated by the betfair cross matching algorithm or whatever it's called.

cross matching doesnt work across markets shirley? when did that come in?

Cross-market as well as cross-selection virtual prices and matching were part of the original plan as outlined by bot. Many moons ago. I've not seen it mentioned much since, but I assume it was phased in long ago. And don't call me shirley!

Fred (I won't call you shirley) there's still lots of cross market arbs available.

Trevh, that's cool. I was obviously looking in the wrong places when I last checked (been quite a while but it was football CS, GS, and U/O etc that I looked at).

Would you go so far as to give an example where cross-market arbs may be witnessed?

I think I may have got hold of the wrong end of the stick.

Are we talking about intra-exchange cross-market arbitrage or traditional arbitrage involving betfair and one or more bookmakers?

think its just different betfair markets, U/O & CS etc

That's what I was thinking brendan. Which type are you referring to john_Nap and Trevh?

cross market arbs are extremely hard to find given that you'll be effectively charged the commission on the gross not the net profit, which would mean you would need a large deviation in prices!

Betfair cross-matching is about matching across different results in the same market as far as I've seen any evidence of.


Personally I find it better to try to work out which of the odds is the 'wrong' price, and use that and then hope that market corrects to the other.

I also think there is a lot of opportunity for looking at prices in the overs markets in football. For example you'll sometimes see odds that suggest over 1.5 is 1.3 and over 2.5 is 2.5 or something, implying a huge (unrealistic) probability of exactly 2 goals. (This also sometimes then disagrees with the correct score market for 2-0,1-1,0-2 combined).

In that case you can dutch under 1.5 and over 2.5, but again the commission on turnover can really hurt this, so you have to be very confident the market is wrong.

Better is to try to arb with a high street* bookie. That way you only get charged commission in one direction, and if you win on betfair you get to pay a decent amount of commission, reducing future PC charges, and if you lose you get a good dose of implied commission, reducing future PC charges, and effectively have made a quick withdrawl from betfair with some profit.

*Internet version of ofc.

Cross matching was originally brought in under the guise of providing liquidity to inter related markets, Fred. Cross market arbs do crop up not sure if thats due to cross matching being turned off at times, but it's hard to find cross market arbs that'll cover comms for most of us.

You've a better memory than I Joe. After reading some posts on here the other year (and forgetting about cross-matching) I wrote a bot that looked for quite a few different cross-market arb scenarios on football. Running over several days it didn't find any arbs that would even cover 2% commission.

Even though it is good for us () - When they first brought it in they "forgot" to tell us about it

Still waiting for a response to my question at the time: If it's best execution - how are you making money from it?

But nothing in my in-box yet

Among other things: Rounding errors. You can quite often squeeze out extra pennies by using odd bet amounts (esp when laying) so you get an extra penny for no extra liability.

What happens to that penny if you don't optimise?

Someone on here once had a chart which said which "side" to bet on for each tick, i.e. which was the pure odds and which got shafted by rounding, but that's going quite far.

i think there is a cross market matching working or the markets due to bots, i guess, are very efficient. i wrote a bot to find opportunities in CS and match odd however the difenrce in odds were not sufficent to cover commission.

Yes ror, also, in a two runner market with cross matching on, you should never bet at odds of 1.04, 1.05, 108 or 1.1, always bet the inverse at 26, 21, 13.5 and 11 and you have a better chance getting a better price.

For instance if you back 1.04, you will get 26 for the inverse. If you lay 26, you might get some 22 or 25 or whatever.

Little improvements at the margin.

Yup, very important when trading the end of cricket games investor, when odds are often that short for long periods.

However, as a £2 staker, I'd often love to lay 22 rather than backing 1.04 but it's way beyond what my balance and (self imposed) liability limits would take, or isn't compatible with keeping any 'green' I've built up.

I usually just abstain when odds get short, but one day I'll be a high volumer I hope. I need more confidence first though, I've ignored some spectacular 1.04 shots recently. 4% really is nothing to be sniffed at and not taking those chances separate the profitable from the grinding even players imo.

The other big reason to take intra-betfair cross-market 'arbs' (even at breakeven) that no one has mentioned is minimising comm.

Having said that, I don't see any pure arbs and will only take a position on 'implied wrong odds' when I have a strong view.

I am pleased to see that you took my advice and posted the same issue here John.  It is ok, you can thank me later.

The fact is that cross market arbs do appear but not with sufficient volume to share them around.  Even if I did find a high volume break even cross market arb then it would be snapped up if I told you.

as investor says, they exist but bots are generally all over them. and people running those bots tend to be on lower commission than the average punter, so unless you're on 2% they can break even at prices which will lose you money.

in other words, if you're hoping to find any sitting there waiting for you you're likely to be disappointed.

Sod the facts. Conjecture is good enough for me. Now where should I point my bots?

Odds     Stake(£)     Bid type     Placed     Profit/loss(£)
23.00     16.00     Lay     30-Nov-11 22:10     16.00
36.26     15.00     Back     30-Nov-11 22:10     -15.00

That was a perfect cross market arb.

^
If that's all I did (manually), I don't think I'd make any decent money though.

Well 3 more like that and you can buy a pint

There are definite algorithms present in market movement and boundaries on the betfair exchanges, It is also noticed that favorites are sent to sync more with industry prices but hardly ever plunge below the competition, cross market trading is much more difficult because of these extra controls, if you observe when middle priced runners have been brought down in price by weight of money and yet don't seem to descend past the industry equivalent price as soon as the race reaches start time and there is a let up in betting amounts these prices can rebound out a number of points as if some mechanism was imparting a bias against the price being pushed in, the effect is clear and quite interesting.

It would be very interesting to know just how much code and what type goes into market controls on a betfair exchanges as knowing the mechanisms can reveal its benefits and achilles heals to bot programmers.

abolo
Date Joined: 24 Jun 11
Add contact | Send message
When: 02 Dec 11 02:11
Joined:
Date Joined: 24 Jun 11
| Topic/replies: 2,922 | Blogger: abolo's blog
Well 3 more like that and you can buy a pint



Well if you did this a thousand times, you could expect to make an average of between £5.85 and £8.65 gross per arb, depending on which price is correct as a lay win wins me loads and the profit from the arb is not evenly green.

Taking 23 as correct price, I win £1.00 95.65% of the time and £176.90 4.35% of the time.

thx for the explanations, i'm sure you're doing fine, it was just for the joke

Yes abolo, cheers!

There are definite algorithms present in market movement and boundaries on the betfair exchanges, It is also noticed that favorites are sent to sync more with industry prices but hardly ever plunge below the competition, cross market trading is much more difficult because of these extra controls, if you observe when middle priced runners have been brought down in price by weight of money and yet don't seem to descend past the industry equivalent price as soon as the race reaches start time and there is a let up in betting amounts these prices can rebound out a number of points as if some mechanism was imparting a bias against the price being pushed in, the effect is clear and quite interesting.
_______________________


i'm not a stupid person but everything u just typed when right over my head

Supersynico, have you heard the expression 'bs baffles brains'?