FGS and AGS ratings (football)

[princecapri]

Howdy do, This is an idea I had last season, and was put into practice at the end of last season. It yielded good (not amazing) results, and I wanted to discuss this here to see what people think. This is merely a discussion, and I want to see what other people think. This is not a fool-proof or properly tested system. The Idea: is to simply pick FGS and AGS bets from football games based on recent history. This proposal thrives on banking on players in form, and not on one-hit wonders, or players who are good for short part of the season. The Ratings: For any game, the points awarded are as follows: First goal scored: 1.5 Penalty scored: 2 Any time goal: 1 Maximum points awarded are 2 (so if someone gets a penalty, which is also the first goal: 2 points. Or AGS which is a penalty, still 2 points). Also, points are given for 10 previous encounters. For the most recent encounter, the points are multiplied by 10, then 9, .... (Example covered later). The Opposition: Rooney playing against Derby has more chance of scoring than Adebayor playing against Chelsea (ideally). Hence if the 'potential number of goals to be scored' is taken into account, each week, a few best bets can be found. The Classification: In order to find good value (Strikers generally have small odds, midfielders have bigger odds, and defenders have massive odds, conventionally), it can pay to back defenders in long run. After looking at some figures for last season (EPL only), I realised that strikers, midfielders and defenders can be considered at par if you further multiply their ratings by a factor of 1, 1.6, 7 respectively. So a goal scored by Sendros is 7 times more valuable than that scored by Adebayor. The Problem: This system relies on extensive list of all scorers and cannot come into affect till 5-10 games have already been played. So for example, I had over 90+ scorers for EPL at the end of last season. Another issue is you cant track each player individually. For example, if Dirk Kuyt doesnt play a game, should he be awarded 0 points? I think yes, since he is 'sitting out' and so is losing form. An Example: Consider 2 players: Berbatov and S Larsson. Their points in last 7 games are: (Latest first) Berb: 1, 0, 1.5, 0, 2, 1.5, 0 (striker) Lars: 0, 1, 1, 0, 0, 1.5, 0 (Midfielder) Their total points (first multiplier is 7, then 6, ...) are: 23.5 and 14 As can be seen here, Berbs is obviously more likely to score than Lars This value can be further offset by using the classification, which means that Berbs will have 23.5 points, and Larsson will have 22.4 points. Introducing classification means that if both players have same odds to be FGS, then (in this case), Berbs will be a better bet. Still, if opposition is also taken into consideration, then a better picture can be seen. For example, if Tottenham play Chelsea, I expect Spurs to score 1 goal. And if Birmingham play Derby, I expect Bham to score 3 goals. This would tilt the balance in Larsson's favor, and would mean that (esp with) odds of 10/1, Larsson is a better bet than Berbs (6/1). The Conclusion: The system is quite simple (although can be made more complicated), and would thrive on betting on in-form players. The only problem, like I cited before, is the extensive listing. This is currently implemeted in Excel, but I will be tempted to move into C++/Java if someone cleverer than myself would be willing to experiment. Please feel to discuss this, as I'd be happy to talk about it further.

[Bedlam]

Re: FGS and AGS ratings (football) It all sounds very logical mate. I made a few bob last season, just by doing doubles from 5 in form strikers to score first goal (not selected nearly as scientifically as your method!). It's well known that confidence breeds confidence, and strikers always have 'purple patches'. Your idea is very interesting!