Friday, 19 August 2011

Profiting from being 'nearly right'

Having paper traded qualifying matches for 4 days, with 92 qualifying matches I think I can start pulling some conclusions together. Below is the summary page from my spreadsheet:
With a strike rate of nearly 36% and a return on investment of just under 25% this little experiment tells me, first and foremost, that my simple selection criteria are actually quite accurate and reliable. The frequency of our scores, in all three instances, is above the averaged frequency figures for the major European leagues (i.e. in our games 1-1 has been the result in 13% of matches against an average of just over 12% in the leagues.). This is just what I would hope and expect to see as we are actively trying to seek these results out!

Whilst it would be interesting to continue with this exercise over a more meaningful time scale I am going to leave that to others more motivated than me! I am convinced that I don't need to tweak my match selection criteria too much to produce a workable mechanical trade, although in reality I often so this trade with favourites a bit shorter and dogs a bit bigger than the limits I set myself here. When I look at some of those 92 games I am convinced I would have made more trading them - the 2-2 matches for example - must, by definition have passed through 1-2 or 2-1, and some of them would have passed through 1-1 as well. The green opportunities were there!

Tuesday, 16 August 2011

What a difference a day makes!

The paper exercise on Scatter Gun set and forget punts (I can't really call it trading, can I?) is looking very interesting.

This screenshot shows the results from the first two days:

After Day one, you'd scrap the idea. After day two you'd remortgage your house! Seriously, for me, two things stand out - firstly yesterday saw an abnormally high incidence of 0-0 scores, and far too few 1-1's ... BUT, secondly, a strike rate of 30% saw a tiny profit.

This screenshot shows the running totals so far:

It's too late and I'm too tired to write more about this tonight, but again look at the return from a modest strike rate.

More tomorrow....

Sunday, 14 August 2011

Set and forget Scatter Guns..a follow up

The idea of trying the set and forget SG's on the opening day of the Premiership was mainly just a bit of fun. There was, however, another factor which made me want to try it. There is a small group of traders, calling themselves the Argy Bargies, who occasionally trade the late night / early morning (UK time) Latin American matches in Brazil, Chile and Argentina, with the occasional foray into the USA and the smaller South American countries. As I do need to sleep I have been using set and forget SG's on the later matches quite a lot recently with positive results, and, as always, that got me wondering...

In last night's post I asked for mathematically minded people to give their thoughts on the outcome and on my analysis of it. This generated an interesting response from Pete, whom I forgive for his gentle jibe at the sporadic nature of my blogging :-), which is well worth a read.

One of the points Pete makes is with regard to the returns from the three scorelines:
your profit over the different outcomes isn't uniform, currently if you get 75% of your winners at 1-1 you'll win far more than if you get 25% of them at 1-1. Had all your winning matches finished 2-1 to the home side you would have finished with a £7.20 profit, had they all been 1-1 you would have made £110. As it is by betting on all 6 games you would have made £62 not the £10 quoted (I believe you averaged rather than summed the games) From Qualifiers only it was £27 profit. 
The reason I stake the SG at multiples of 5 and 2 if trading the game is because the key to the SG as a trade is the 1-1 scoreline, so I want a larger profit on that to provide me with options after the first goal is scored. As a set and forget, would it better to adjust these stakes? Dutching them, to give an identical profit across all three is an option, and on the average odds used in the Prem games would have given an even profit of £23 bar a few coppers. That would have resulted in less profit than was actually generated (£42 against £62).

Some basic research on www.soccerstats.com confirms that in all the major leagues 1-1 and 2-1 are much more frequent results than 1-2 so after playing around a bit with the stakes I have concluded that a better way to stake this as a set and forget would be £4 on 1-1, £3 on 2-1 and £2 on 1-2 (or multiples thereof).

Pete also refers to the issue of match selection:
I think the main issue before you can contemplate an edge is game selection obviously blindly backing 1-1, 2-1 and 1-2 in either every game or game with teams who are close in the markets isn't going to work. As you're rulling out straight away games like man utd on the road where they do tend to draw alot (50% of matches last season) yet will likely be an odds on favourite.
Hard to argue against that! When trading the SG, particularly in the top leagues, I would rarely ignore some  odds on favourites. In the prem, I wouldn't ignore ANY of them - the league is so competitive now that 1-1 is frequently hit at some point during the match and provides options even if the final score represents a rout!

With my trading hat on, when I first started playing around developing this, I researched the games thoroughly. I looked at the form, head to heads and the prices on M/O, CS and the goals markets on Betfair prior to deciding whether or not to enter a trade. I rarely bother now! Using such small stakes together with various other enhancements (for example scalping some 0-0 cover pre match and entering the U2.5 goals market immediately after a goal to get cover - live example as I type, Kilmarnock 1-0 against Hibs after 3 minutes - entering U2.5 at 3's) means this trade is so generally successful that I can afford the odd wipeout.

I am now confident and comfortable enough with the SG that given relatively evenly matched teams (even if I've never heard of them or know little about their respective recent form, league positions, team selection or anything else for that matter) and decent liquidity I will enter a trade.

Going on from here this is what I'm going to do, again as a paper exercise. With neither team below 2 or above 4 in match odds I will paper 'punt' the SG scorelines. I'm not patient or motivated enough to look up the CS odds so will use 7.2 for 1-1, 12 for 1-2 and 10 for 2-1 as the odds, staked £4 on 1-1, £3 on 2-1 and £2 on 1-2. Just for fun..... unless the results start suggesting otherwise....

As an aside I entered U2.5 in the Killie game at 3 and got out at 2.84. It is now 1-1, after 18 minutes so I've exited the trade for a small loss on 1-1 and small greens on the other two. The goals came too early in this instance, but a profit ensues. 1.01 the bloody game ends 1-1 now!