Prediction Analysis

⚡

Continuing on in this series of analyzing the predictions given by AI I moved onto NCAA Basketball

Once again my hypothesis is that there must be a correlation between the predicted scorelines given by AI and the actual games totals, and by finding these triggers we can find new, optimal ways to bet

I found clear correlations for soccer and now I’ve analyzed the last week of college basketball predictions in an attempt to find the patterns

Here are my findings:

I found a clear trigger for games to be high scoring

Everytime in the last week the AI gave one team a predicted score of 80+ that game comfortably covered the over

There were only 5 occasions in the last week when this happened but here are the totals in those games:

• 165

• 160

• 150

• 178

• 138

The last of those results came last night, and I sent a message to paid subscribers before the game so they could look out for it

The prediction was Houston 82-76 Iowa State

The 82 served as the trigger to bet the over

Bookmakers had the total line set at 129. So as a test of the waters I selected the highest total Bet365 offered…

This is a positive finding that should give you an extra angle to read the tables with. When you see the 80+ that’s your trigger

Conversely I pondered that if we have proved that predictions with 80’s imply overs then it would be safe to assume that there is a trigger for low scoring games, and last night there was also a perfect example of that

Along with the Houston prediction users also got this:

Virginia Tech 63-68 Virginia

Although the moneyline prediction was incorrect, the prediction of both teams in the 60’s caught my attention

The bookmakers had the total line set at over/under 126.5 points

The game ended with a total of 116. Confirming my suspicion that when you see two teams predicted in the 60’s the game will be low scoring

The 80’s and double 60’s predictions seem to be rare however when they do appear they have had 100% accuracy

The only other commonality I found was when both teams are predicted in the 70’s and the score is within 6 points (i.e 70-75) then it beats the over

I’ll continue to look for these patterns as we move forward, and I hope these findings can bring more clarity to the AI predictions and arm you with new ways to create bets using the tables

For the next prediction analysis I will be dissecting NBA

NoNonsense.