historical data and value


Historical data and value

I think am very different to a lot of traders in that I don't look too far back for trend indicators. In my purely personal experience, there is very little to be gained from mining historical data from a period where the factors that determine the outcome were significantly different.

Now clearly that can vary to an extent from club to club, but certainly I define particular break points in historical data. The top three:

1. Season. Once the new season is underway, I exclude previous seasons' data. My logic is that this clears my mind of a large number of imponderables and allows me to focus on the more recent, and imo, more relevant statistics. 

2. Personnel. Change of manager, significant change of playing personnel which can lead to changes in style and/or game characteristics. Simple example, Liverpool lose Suarez and Sturridge gets injured, ergo they will be less dynamic in terms of goals, chances and style of play. 

3. Head to Head. I appreciate some may find value in studying this. I see none. Head to head stats in the main are no different to tossing a coin. The outcome of games (outside recent data) has no more relevance if the game was between these two sides than If it was against another random opponent.

Whereas I might look at the bigger picture to find longer term general trends ( goals per game average, ratio of SOT to goals, structure of game re: timings of goals, HT/FT, possession: result ratios etc.) I see three months or the beginning of the season as the starting point.

For me this is enough to determine if there is a trend or just a statistical variation. Think of it like you would think of fluctuations in price when pre match trading. Individual results may well form part of the normal variance, ie most sides will lose occasionally without that suggesting any greater overall trend, but in other circumstances you see the trend "breaking out" from the current bands, at which point they become of great significance.

Again, a three month window tends to be enough for me. In that same context, I certainly follow the league position trends. A side which is dropping consistently down the ladder is evidently performing significantly below par for the competition. That, in my simple mind, is a good indicator of how they are going.

Everything else I ignore. "They never win away in London" " They play better on a Wednesday" " They beat more teams in red shirts than blue " Even if they are significant, I don't have the brain power to factor these things in. So I am better to ignore them.

I think that punters, and traders in particular, often fixate on the minutiae of statistical analysis and fail to clear their minds and look at the bigger picture. It's a case of not seeing the wood for the trees, if you get my drift.

I accept again that those who are looking at creating statistical models or working with neural algorithms and the like will need to harvest vast amounts of data to do so, but I remain a bit old fashioned in that regard. I also feel that we are in danger of looking for the non existent holy grail. The size liquidity and intelligence of the markets mean that the statistical probability of an outcome occurring will be pretty accurate BASED ON STATISTICAL ANALYSIS.

Therefore, I think there is a danger of wasting energy and effort panning for gold in the statistical streams when we should, by rights, simply end up proving that the price is right. Better we look at reasons to feel that the trend will be bucked. Reasons to enter contrarian trades - even if the logic is difficult to prove.

Without being arrogant, I have been consistently profitable over a number of years. I attribute that to keeping my analysis simple, having clear entry and exit strategies and doing the simple stuff consistently well. The same scalping opportunities, the same odds bounces, the same price movement characteristics happen week in, week out. Learning to identify those is more important than forensic analysis of stats. For me. 

All of which comes with a caveat. I work differently to many other traders. They are probably better to follow than me, but I still stick to what works for me. I try to be very selective. I only really like to trade on games I can physically watch. I only like to trade on competitions I am familiar with. That is my comfort zone and whenever I move outside that, my performance suffers. It happened last month. That perhaps means that my edge is less scientific. More hunch based or whatever. I try not to over analyse that either. That's also why I am not very good at posting specific trades. A lot of my work involves creating a good position through scalping and pre match trading which makes leading a trade difficult. 

1 comment

Leave a comment

Get in touch