Let's start by explaining what the score predictions are NOT.
They are NOT a magic bullet that will lead you to the the land of riches. You cannot assume that you can randomly pick games from the list each week and come out ahead in the long run. You cannot only pick games you're interested in watching, take the score prediction for that game and run with it, hoping to come out ahead.
The score predictions are a HANDICAPPING TOOL. A tool to be used in conjunction with other handicapping methods and common sense.
Can the score predictions be used as the only source to pick games each week?? Well, yes. I do not use them that way. However, if you follow some strict guidelines, you could use just the score predictions to come up with plays each week. However, if you do so, you MUST MAKE SURE YOU ACCOUNT FOR INJURIES, WEATHER AND OTHER VARIABLES SUCH AS MOTIVATION, that can impact the final score of a football game.
If two offensive lineman go down, or a QB goes down or anything along those lines, it will obviously have an impact on things going forward and you have to take that into consideration.
You'll be most successful using this model, if you already have a handicapping method or way of going about things in place and you use this in conjunction with your methods. There are guys using this model that are simply phenomenal handicappers. This is just another weapon in their arsenal.
But I do understand there are some less experienced bettors on board. So, I'll try and offer some guidelines with you in mind.
Ok, on to the games............
The model has always been strongest in two areas.
1) Underdogs it picks to win straight up, or come close, ie the model has them losing by a point or two.
2) Favorites it predicts to DESTROY it's opponent.
If someone were looking to use only this model for their plays, I would advise them to use it for underdogs that the model picks to win straight up or lose by a point or two. Those have been money in the bank over the years.
For example, a team is favored by 7 but the model has them losing by 1. In this scenario, I look at the underdog very closely. I'd need to find a very strong reason NOT to play that dog. Likewise if the model has that 7 point fav winning by 1,2,3 even 4 points. I'd still strongly consider grabbing the +7 in that scenario.
I'd look at underdogs of +3 or higher in the above scenario.
The other scenario that has been profitable is large favorites, which of course are very common in college football. For example, a team is favored by 20 and the model has them winning by 30 points or more. More times than not, the model is correct in those scenarios.
I'd look at any fav of -14 or higher that the model has winning by much more than that.
Those are the two scenarios to consider if using the model as a SOLE method to pick games.
If you're using the model in conjunction with your own methods, ANY GAME WHERE THERE IS A SIGNIFICANT DIFFERENCE BETWEEN THE POSTED LINE AND THE SCORE PREDICTIONS IS WORTH A LOOK.
Lastly, the model is also particularly strong in NON CONFERENCE GAMES, which also makes it valuable during bowl season. This is because it's easier for oddsmakers to post a line on a conference game, especially late in the season. By that time, you're dealing with teams that have played schedules similar in strength. However, that's not the case with non conference games and the model does a terrific job in analyzing the difference in schedules and applying it to the predicted score.
If you see an 0-4 team from a tougher conference playing a 4-0 team from a weaker one and the model has the 0-4 team winning or covering, put faith in the model. It will sniff out the pretender. (another area that makes the model priceless in college hoops as well).
EARLY SEASON - The predictions will grow stronger each week. However, the first couple of weeks we are forced to use LAST YEARS NUMBERS.
For example, we have the scores for week 1 college ready to go. I will send them out shortly. But those scores are using last years stats with some tweaks. For example, we have filtered out early season results from a year ago and have also filtered out the last week of the NFL season to improve accuracy.
You can still use the model the first couple of weeks, but you have to use common sense. If two teams are playing each other and are largely intact at key positions, use the model. If two teams are playing each other and only return 5 starters on each side of the ball and a QB is missing, DON'T USE THE MODEL.
Around week 3 or 4 we'll be using strictly this years stats. From that point forward, the model will get stronger each week as more data gets pumped in from the current season.
Also, as we hit the midway point or so, we'll start posting 2 sets of predictions, one using full season data and the other using only the last 4 games. This is crucial come bowl/playoff time as it can show teams fading or getting hot at just the right time.
Lastly, the model can also be used for TOTALS. The larger the difference in the predicted total, the stronger the play.
Hopefully this guide is helpful. Of any product we have offered over the last 20 years at Bettorsworld, the score predictions have the highest number of subscribers returning each year since we have offered them.
Enjoy and Good Luck!
P.S. Sportsbetting is a GRIND. You are realistically shooting for 54% to 58% winners and in all but a very few instances, the number of a consistent winner long term will be much closer to 54% than it will be to 58%. That is a FACT, regardless of what you may read on the wonderful world wide web.
Over the course of the last 30 years, I have taken action, I have moved games for syndicates, I have bet professionally, I have operated sports monitors and come in contact with some of the absolute SHARPEST MINDS in the sportsbetting world. 60% long term does not exist.
If anyone thinks it does, I'll gladly put up a significant amount of money that says that it doesn't. I'll win that bet. No question about it.
Keep that in mind when using this tool. This can help you get to 54% from below 53%. It can't get you to 60% from anywhere.