Attempting To Pick The Over Under In A Sport Game

I'm having a debate with my buddies regarding the Over Under in a sport game (ie. NHL, MLB, NFL etc).

If I was going to attempt to pick if the game to be an over or under, how should I approach the analysis of both teams in their past performance in the same season?

Some of us has stated that we should analyze the Mean Statistic of goals of both teams ONLY playing against each other regardless if the team was playing home or not.. Then cross reference that Mean statistic value against the upcoming game to pick the over under from it.

Others of us has stated that we should analyze the Mean Statistic of goals of both teams playing against ALL teams in their league and cross reference the Mean statistic value against that with the upcoming game to pick the over or under from it. 

What do you guys think.  Is this a good way to attempt to pick the over under in a game or what additional methods do you guys find works the best for you..

thanks!

tom

I missed the 'over/under' portion of the workshop, but I was told in general that using one of the above methods against a specific statistical formula will yield a higher winning picking percentage over time as given that there will be some games that teams will have an off day or a complete blow out.   He validated his workshop by showing how he was able to successfully pick the team win, the over under and the point spread for 7 years in a row against the NHL and NFL leagues alone.

Coin flipping is the easiest, but 'big data' is just as important.

https://www.metacafe.com/watch/an-ajHHbJJtYh7n2/fast_times_at_ridgemont_high_1982_brad_and_lisa/

You sound like a single successful guy, so I'll let you in on a little secret.....

It's ketchup & mayonnaise. 

Also, coaching and schemes vary from league to league.  IMO the coaching staff and the system they've implemented will obviously have a bigger impact in a football/basketball game than hockey/baseball.

Bottom line is that it requires much more than resarching avg scores to be successful at picking totals (by successful I mean >60%).  That being said, its entirely possible to pick at or slightly above 50% using your system or really any methodology for that matter becasue at the end of the day it is essentially a coinflip.

Cause books ONLY have less than 1 minute to post the total for 2nd half ( when the 1st half was end )

Total for FG , or 1st Half : You "think " you got the INFO ? nah...

BOOKS know it too ( rain, windy, snow,etc...) . That's why  they  " set " those numbers 3-5 days in advance ... for you to bet (on) .

GL,