[omedo]

I'm facing this question myself right now.
And i want to know the opinion on the more experienced skilledgamblers around here.

Obvius the likelihood of a team winning the game is heavy correlated whit his likelihood of scoring.

But for someone starting where should be his focus?
Is easier to sart on the scores?
Is easier to start on the capacity of winning?
Or you just cant dissociate one from another, even in the beginning?

[Snowball]

Play what you know.

The sports you'll do best at are the ones you understand the best
or have the most experience in.

Play what you know.

Sure, some people will dispute this, and say it's all about probability, stats, math, etc..

but you did ask for an opinion.

[Waterstpub87]

I'm facing this question myself right now.
And i want to know the opinion on the more experienced skilledgamblers around here.

Obvius the likelihood of a team winning the game is heavy correlated whit his likelihood of scoring.

But for someone starting where should be his focus?
Is easier to sart on the scores?
Is easier to start on the capacity of winning?
Or you just cant dissociate one from another, even in the beginning?

You focus on creating a good fundamental model that projects a final score for a game. This will allow you to create your own spread and total, which can then be compared to the offered odds. You can then use this to calculate your expected advantage.

[TheMoneyShot]

I'm facing this question myself right now.
And i want to know the opinion on the more experienced skilledgamblers around here.

Obvius the likelihood of a team winning the game is heavy correlated whit his likelihood of scoring.

But for someone starting where should be his focus?
Is easier to sart on the scores?
Is easier to start on the capacity of winning?
Or you just cant dissociate one from another, even in the beginning?

In my opinion... I've been wagering for close to 20 years.... I don't see anything consistent in regards to Totals. Nothing of consistent value.

I think you should just study opening lines on games.... and watch the closing line. Get a feel for what's going on.... why a particular side won or lost. I have much more success on this.... rather than wagering on totals.

[HeeeHAWWWW]

In my opinion... I've been wagering for close to 20 years.... I don't see anything consistent in regards to Totals. Nothing of consistent value.

There's a reason: totals betting is inherently more difficult than ml or hc. Same factors plus some more, and combined in more complex ways.

On top of that, in a lot of sports you get worse juice on it as well :-)

[Optional]

Totals tougher but for some reason appeal to a lot of new gamblers.

[Rich Boy]

Totals markets are always easier to beat, just look at the limits and volatility in the lines.

That alone tells you how inefficient they are.

Focus on totals for smaller markets and you are more likely to have an edge (College, NHL, MLB)

[gojetsgomoxies]

as mentioned, play stuff you know..... and play either point spreads or ML of fairly even games. that's the simplest way to start.

i'd also look at power ratings alot in the beginning. you can learn alot in general just from looking at them alot i.e. they summarize alot of info and will help you get up to speed with the teams even if they aren't terribly helpful directly in being an ATS winner.

[evo34]

Totals are not tougher. A $500 bet from a marginal-edge player can move a CBB total a full point at Pinnacle. That's how unconfident they are about their numbers. My only advice: look at totals in conjunction with the restedness of teams and the time of season. Offense/defense are affected quite differently by fatigue in basketball and hockey.

[TA3]

I'm facing this question myself right now.
And i want to know the opinion on the more experienced skilledgamblers around here.

Obvius the likelihood of a team winning the game is heavy correlated whit his likelihood of scoring.

But for someone starting where should be his focus?
Is easier to sart on the scores?
Is easier to start on the capacity of winning?
Or you just cant dissociate one from another, even in the beginning?

I'm late to this thread - so apologies for dragging it back up if you are set as I've been preoccupied with my soccer prediction model and was over in that forum since NFL was basically over - but here's what I've figured out from modeling the last several years of games.

I developed an algorithm which produced a 5-year average win rate of 60% with a 14% ROI. Swings were not wild but in 2012 I had a losing year of -7% ROI.

I did this by only using the following information:

1. Score for each team in each game (that means how much each team scored and allowed and recognition of home vs. away)

That's all I used. There may be more complex ways of doing it, but the results I found were fine by me. However, there are two key things that need to be done with this data to make it produce something of value because what I've described already can be done on the back of an envelope...

First, you have to figure out what to do with that small dataset. Better to refine what you do with that dataset than add fields and fields of extraneous info for little benefit and, worse, negative benefit. This is where the trick is of course, but I wanted to at least try to focus your efforts so you aren't swinging wildly with no forward momentum. Ultimately, I found that the number of recent games I used in my dataset was critical and the way in which I compared possible scorelines was critical. I ultimately came up with several methods for projecting scores and used them all.

Second, you need to develop a set of rules by which games which are too close to call are eliminated from consideration. My analysis always ruled out half of each week's games because they were just too close. I compared the results from each method determined which of the methods - when they were in agreement with other methods - produced the most wins. This allowed me to rule out games where methods x,y,z didn't line up. This is critical. You have to remove games from the pool that are suspect - a removed loss is as good as a win here.

Final thought - most people go wrong in their modeling by factoring in full season stats and then testing. You can't do that as it gives false results. You have to model each game as if you only have knowledge of scores from games played prior to the game you are modeling. If you don't do this your results are flawed.

Good luck. Hope this helped. Once NFL rolls back around I'm going to be more active on this side of the forum so we can compare notes then. See ya in a few months.

Here's a graph of how my results moved during a five-year period. I backtested 10 years and the results were similar - 3 losing years out of 10 with a 57% win rate and 9% ROI for the 10-year period. All assumed -110 odds.

[omedo]

thank you all for the ideas here explained.
More are welcome of course.