Totals

[ATX]

I have a lot, if not all of them, but they are all hard copy, I lost my db a while back and now have everything on paper.

I believe donbest.com has totals with line moves, go to the archives section. To go back further than 9 months, you can manipulate the date which is in the URL.

If you can stomach typing his name, JimFeist.com has a database, I think. I havent been there before, but I read this from Big Lou. Correct me if I'm wrong about these.

 

[SixthSense]

Year Overs Unders Pushes %
1983 113 107 4 51.4%
1984 104 116 4 47.3%
1985 112 108 4 50.9%
1986 95 126 3 43.0%
1987 86 83 41 50.9%
1988 105 113 6 48.2%
1989 108 114 2 48.6%
1990 107 112 5 48.9%
1991 101 119 4 45.9%
1992 103 118 3 46.6%
1993 105 114 5 47.9%
1994 109 111 4 49.5%
1995 131 106 3 55.3%
1996 116 118 6 49.6%
1997 114 122 4 48.3%
1998 120 115 5 51.1%
1999 121 122 5 49.8%
2000 117 127 4 48.0%
2001 119 127 2 48.4%
2002 128 124 4 50.8%
TOTAL 2214 2302 118 49.0%

 

[SaintJoe57]

Thnks it would seem to me that you'd have better luck just flipping the perverbial coin!

 

[SixthSense]

SJ,

Yeah, there are situations within those that work, but as a whole, the linemaker does a nice job of keeping the playing field level. Also, I believe,in weeks two and three, the unders dominate more so than in other week.

 

[viggme]

I've never understood the reply "both teams must cooperate" in regards to playing totals. You must handicap both teams when playing sides or totals.

JMO sides are a little sharper. Totals are easier to beat, especially openers.

 

[Potter]

If you're going to take a look at totals,you've got to look in terms of scoring events(TD's FG's) instead of actual points on the board.

If two teams post 21 points(around the league average per team)one could have done it with 3 normally converted TD's,while the other could have needed 7 FG's.

One busted play in each game could respectively turn a TD into an interception(21 now becomes 14) & a FG into a TD(21 now becomes 25).

A huge swing on just two plays.So in terms of totals the bare points can lie.

P.

 

[acw]

Potter,

Have you been working on totals for European football too?

Personally I go as far as working out the chances of the exact scores for every match, so the Over/Under 2.5 goals total runs out of it kind of automatically.

 

[Potter]

Hi acw,
Soccer totals are probably easier to model(both for bettor & bookie)because goals follow the Poisson Distribution fairly well.

And then as you say you can work out probabilities for all the potential correct scores & then work out your odds for the totals from there.

However the results are dependent upon you having an accurate & predictive value for a team's average goalscoring(& conceding)potential.

An exponentially smoothed time series is the usual method used to produce these.

The next problem(that's common to both the Nfl & soccer) is to decide how a team that scores for example 1.5 goals/game against an average team will fare against todays opposition.

They way to go here is to calculate the team's rate of scoring compared to the average for the league.

Say Oakland average 3.3 TD's per game & the league average per team is only 2.4.You can say that the Raiders score 1.375 times more TD's than average for the league.
If you do the same & find the rate at which their opponents concede compared to the league average,(& then do the same for the FG match up)then imo you're on the way to getting a more realistic value for both total points & handicap lines.

If you just use average scores your predictions cluster around the average & you hardly ever get to predict that Chicago will struggle to get out of single figures against a team like Tampa.

P.

 

[acw]

An exponentially smoothed time series is the usual method used to produce these.
What I like about that method is that it also gives valid predictions in extreme cases.

They way to go here is to calculate the team's rate of scoring compared to the average for the league.
Personally I try to work on a system without splitting the teams up in different leagues. Effectively I want to be able without too many variables to predict results for any team (including a mixture of international and national teams) playing against each other. It may not work out, as teams like Ajax and Celtic in their leagues (even taking into consideration their strength compared to the others) score many more goals than when they would play against each other.

 

[RUSS]

do you post your totals?

 

[Potter]

acw,
hadn't really thought of one universal ratings system for teams,it would certainly make things easier

I tend to stick to ratings within leagues because I think that the character of the individual league has a major influence(possibly the major influence).Team's may seem to be showing very unusual(short term)behaviour but eventually they are dragged back towards what is considered the league norm.

You can usually accommodate new arrivals into a league by looking at how relegated or promoted teams of a similar standard performed in the past.

With different country European match up's there's similarly a fairly large amount of historical data to work with.You've just got to decide if the Spanish league is currently superior to the English Prem....& by how much.

For internationals,I know that the FIFA rankings get slated,but allowing for the more obvious descrepancies,you can derive some reasonable predictive regression equations that relate supremacy etc to FIFA rating.

As for the odds on offer on totals markets,certainly in soccer,the size of the vig/overround(or whatever you chose to call it) usually means you need a move by the books in response to punter's money before you get many value opportunities.

I'll try to post some soccer/nfl totals here when the seasons are up & running.

Cheers,P.

 

Potter,

I use the Poisson Distribution for soccer and hockey (sides and totals), but I'm not sure does it work with NFL totas. I use Poisson in baseball (sides, but not totals) also. In soccer and hockey you get 1 point every time you score, but in baseball a homerun gives different amount of runs. Same thing in NFL. I don't know much about NFL, but I think a team can score with these numbers 1,2,3,6 and 7. Your method might work on sides, but I'm very doubtful about totals, but as I said I don't know much about the NFL. How long history you have using Poisson in NFL?

 

[Potter]

Hi Faceless,

I think you're making the same point that I tried to make.The number on the scoreboard is less important than individual number of times that a team scored.

3 TD's & 1 FG is much more significant than the 24 (or maybe 21 or 22 or 23 or 25 or 26 or 27) points that appear on the score sheet.

There's loads of approaches that you can take & seeing if scoring events follow a Poisson is just one.

Although the really important estimate in the NFL is the expectancies for TD's & FG's for each side in the particular match up....unless they follow us Europeans & start betting on correct scores .

I'll post a simplified outline of a NFL totals method later.

Cheers for now,P.

 

[acw]

Potter,

I'll try to post some soccer/nfl totals here when the seasons are up & running.
Could you start posting them on the English Leagues? I am mainly interested in the totals that you would use against the spread firms. At least those are the ones that I am willing to give. Please keep the Over/Under chances a secret!

One more question: Have you ever made any use of forward/backward routines?
Imagine 4 equally strong teams in a league, team A,B,C and D.
Team A plays against team B and wins 1-0.
The rating of A goes up a little and B goes down.
Team C draws against team D. No change in ratings.
Now team A plays against team C and wins 10-0.
The rating of team A goes up tremendous and the rating of team C goes down big time, but also the rating of team B goes up, as it had played against a much stronger team A and the rating of team D goes down, as it had played against a much weaker team C.

 

[Gridster]

"I've never understood the reply "both teams must cooperate" in regards to playing totals. You must handicap both teams when playing sides or totals."

Both teams must cooperate when playing OVERS. You generally need scoring from both teams, and we know from experience that a lot of the time one team doesn't show up. It doesn't help to get 35 points from one team while the other contributes only 3. On a side, you expect to hit if both teams show up, but if only your side plays that day you'll still get out with the hit. Not so in totals.

 

[5teamparlay]

Over Kings 246 tonight

book it !!