So, I've been working on a football simulator for a few weeks now. It's using a giant database that has data on most leagues in the world, teams, players and players' attributes. When I say a "giant" database, I mean it. Anyway, lately I've been simulating World Cup games etc. and I've come to realize that my simulator is extremely accurate so far when it comes to predicting if a game is going to end "over" or "under". For those who don't know what "over" and "under" is, over means 3 or more goals total in a game, and under means 2 or less goals. So far, my simulator has been accurate 100% for 5 games (4 WC, 1 Swedish Cup game). I've only tested it for 5 games yet because I just started. My simulator's predictions were following games were "over" and they all ended "over" (mind you for all these games "over" was paying more than "under", which means these games ending as "over" was less likely): Germany 4:1 England Argentina 3:1 Mexico Netherlands 2:1 Slovakia Göteborg 3:3 Kalmar Brazil 3:0 Chile I'm not yet convinced of my simulator's abilities, obviously 5 games isn't enough and it might just be luck. However, I'm gonna keep testing it and meanwhile I'll release my predictions before games here so we can all see how it's gonna work out. I will only list bets that my simulator gives a 54% or more on. So here are my predictions for today: Paraguay - Japan: Over (60%) Haka - Inter Turku: Over (60%) Lahti - Honka: Under (68%) On top of these, here are some other interesting things. According to my simulator: Spain - Portugal: For this one, 48% Spain, %32 Portugal and 20% draw. My simulator also gave very close results regarding over-under, under was 52% while over was %48. However, the betting site I use was giving 1.35 to 1 for under whereas it was giving 2.15 to 1 for over, so %48 is pretty darn good of a risk to take. Lahti - Honka: Lahti is gonna win (48%) or it's gonna be a draw (40%). Honka's winning chance is only 12% (mind you they're the favorites in this game).