Discussion about form evaluation

[kreeshak]

Coming to the conclusion that form is major in evaluating a game I am here to ask/discuss on how you evaluate form. a) One option would be to measure the points each team has gained through the last 5 games. Problems is that points aren't an indication of achievement. If barcelona gets one point versus a weak team it should be evaluated as a bad result instead of 1 point. However the weaker teams 1 point should be evaluated as a great achievement. So by measuring points we don't get a realistic evaluation. b) Another option would be to measure goals scored and conceded in the last 5 games. Scoring is a sense of achievement and conceding a sense of weakness. The problem here is the noises. If a big team wins 5-0 in one game and then draws the remaining for 0-0 can we conclude that the 5-0 goal difference is a good thing or bad thing? Also winning 1-0 against real madrid is more significant than winning 1-0 against a division three team. c) Another option I was thinking would be to calculate past asian handicaps. For example if a Big team wins 4-0 that handicap would add a big odd to the rating, while if that big team wins 1-0 its less of an achievement and a lower rating would support that. The problem here is that a horrible team that keeps losing would have unrealistic ratings with just ONE win or upset in between failures. Another problem is that a proffesional team that wins every game marginally would have a low rating, while they are clearly do things right even if they just win 1-0 every game. d) Finally I was thinking to calculate ratings given past scores. Real madrid - lesser club. Each real goal is rated 1.40 while each lesser club goal is rated 3.20 (the ratings for each goal are calculated via equivelant odds matches, using prices for first team to score). So if real madrid wins 3-0 they would add a rating of 4.20. While if they win 3-1 they would just add 4.20-3.20 = 1 rating. The problem with this case is that if an underdog scores 2 goals but still lose the rating would be unrealistic. The big question is, how would we accurately calculate form?

[zbrochu]

Re: Discussion about form evaluation I think first two rules are too simple. Bookmakers are extremely good at modeling games, and it would not be enough to beat their numbers. Third rule is a nice one and I know it is frequently used for USSports. Americans sometimes say "Miami is in form becuase they are 5-1 ATS". ATS meaning "against the spread". I read Lem Banker's Book of Sportsbetting last weekend and I know he uses it a lot. And he is (or was? i do not know if he is still alive) one of the best. However he does not have any numerical approach for this, he is just very experienced and have very good intuiton. We would need to create the database that includes ATS results in football and see if there are any corelations there. This would be big effort but might be worthwile.