Thursday, 17 June 2010


Anonymous(2) writes

So basically you're saying you aim for a minimum edge of about 0.5%?? (i.e. taking, after commission, 2.01 on 1 2.00 shot)
No – I am saying that I look for ‘at least’ 2.06 if the market is a 5% commission market. That’s my starting point.

Understand that punts are a very small percentage of my action. Whereas I often risk, at least for a time, four figures while trading in-play, I seldom put more than ₤100 on a punt, and certainly not that much if the value is borderline.

As best I can, I use Kelly on my punts, so that the bigger my perceived edge, the higher the stake and so if I am on at say 2.06 on a 2.0 shot, it’s for little money and if the edge (re-calculated after each trial) turns out not to be an edge, the price I will take next time adjusts.

He adds:
That's crazy. There's no way you could price up markets accurately enough to make your perceived 0.5% edge a reality. You would be having non-value bets an awful lot of the time.

Yes, that would be crazy, but in actuality, and I’m using what I do with the Elo football ratings as an example here, when I price something up at 2.0, and check the prices available, I find one of the following.

a) The price is several ticks higher than 2.06, in which case (after due diligence) I back it.

b) The price is less than 2.06 in which case I drop it.

c) The price is 2.06 or slightly more. For these I typically put in a back at 2.1 and hope it gets matched. On a weekend with a busy schedule, if it stays unmatched, that’s fine. There are plenty of better value bets around. However, if it’s a quiet Saturday, or a midweek game on a quiet night with little action, I may accept the 2.06 close to kick-off just to have an interest for a small amount. Even if the value is tight, I still perceive that I have value.
I'm not a trader but ROI can be worked out easily enough. On the rare occasions where I do move in and out of a market in running, I use my stake as my maximum exposure at any given time. You should be able to do the same.
That’s an interesting idea to use the maximum exposure to calculate the ROI, although I’m not sure that figure is readily available after the event, and one big trade skews the numbers significantly. I suspect that I move in and out of markets a little more frequently than yourself!

I was looking at the numbers from a typical NBA game last night, for which Betfair do actually total up the backs and lays for you, so it would be easy enough to calculate ROI based on the total backed.

There would be significant differences in the results obtained using the two methods though.

Twenty trades each of ₤1,000 and twenty lays of ₤990 would result in a gross profit of ₤200, and an ROI of 1% based on the total backed.

However, using the maximum exposure method, if one trade is for ₤10,000 and the others for ₤10, with a resulting profit of the same ₤200, the ROI comes out at 2%.

Make the maximum exposure ₤1,000 as in the first example, and a profit of ₤200 raises the ROI to 20%.

It would be interesting to know if other traders calculate their ROI and how.


Andy said...

Hi Cassini,

I'd like to invite you and your readers to look at my horse race betting system that I started running recently. It differs from some systems in that it selects horses that are considered to be over priced, so it is useful for people who want to back to win or for traders also. The selections are posted live to my website and I report the results and progress on my blog

I have put up links to your website on both of my sites. Would you mind putting up links to my sites on your blog?

Any comments or suggestions would be welcome. I started the system at the start of the year and so far it has produced consistent profits of 10%.


Curly said...

For a few years I used to calculate ROI whilst I still traded fulltime, now that I no longer do and returned to the world of work I also add an element of the cost of my time which I could be using for potentially better things.

I used to calculate it as my maximum exposure at any one time. So if, using your numbers, I was trading with blocks of £1,000 then my investment would have been £1,000 although if I had taken 3 bites at the cherry before selling any back (or visa-versa) then it would have been £3,000.

I did calculate ROT too which, again with your numbers, would have been 20 backs of £1,000 and 20 lays of £990 for a total turnover of £39,800. But I never considered this a particularly useful figure other than to look at the total turned over and massage my ego a little bit. It was however a simple number to obtain.

The first one isn't an easy number to derive and does involve having to remember your trades which isn't always easy but was never too hard, especially if I had taken more than 5 or 6 unreplied stabs at a price.

Enjoying the frequency of posts as of late, and fortunately quality doesn't seem to be substituted for quantity. Keep it up.

Anonymous said...

I don't quite understand your reasoning on the first point. 2.06 at 5% is effectively 2.01 so you have confirmed you are prepared to bet at no more than 0.5% value which doesn't sit comfortably with me. 2.01 is the equivalent of 49.8%. I don't see how you can expect to differentiate between a true 50% shot and a 49.8% shot although, personally, I do price markets to 0.5%s so may make something 49.5%.
Although not a trader, I'm not sure of the wisdom of being prepared to bet 10x the amount on what you are looking to trade compared to when you have a straight bet and are not looking not to trade. You'll often have to take a big hit, sometimes even the full £1k. And if you operate to small margins (like the 2.06 on 50%) your starting positions are, relatively, too large.

(unless by punt, you just mean an interest bet in which case I have no real issue. But that's not how it read to me)