"Stay Away" Games

Each season, there are going to be games that we find it best to avoid altogether. Rather than rely on anything less than the typical level of accuracy of playing rating data, we avoid the game completely. The system we have takes a great many game related factors into predicting the spread and a big part of that has to do with player ratings. Each player in our system is carefully rated in the form of their contribution to the point spread. So when a starter is injured, the system subtracts their rating value and then in turn adds in the rating value of their backup. Straightforward enough mathematically, but it depends on both the timeliness and the accuracy of the data.

The timeliness issue is solved by constantly running processes that are pulling in data from NFL injury reports. The designation of "Questionable" is frustrating, but is usually resolved right before kickoff, where the player's status is changed to "Active" or "Out". If there are too many players of "Questionable" status that could impact our prediction, a wager recommendation is delayed for as long as possible with the expectation that the injury designations become more clear. In the rare times they do not, the wager recommendation will not be recorded and published on the site.

Even more challenging is the second issue - that of data we are not entirely sure of. Our ratings on all of the starting QBs in the league are dead on, given the volume of data we have to work with to come up with that rating. This is not the case for the majority of backup QBs in the league. Also, regardless of the rating we assign a backup QB, there is an inherent volatility in having the backup step into the starter's role and with only a week of first team reps, expecting a predictable outcome. Too often, a game that has a spread that is within a TD can turn into a blowout when a backup QB is not entirely ready for the challenge.

There are also other scenarios we see toward the end of a season, when playoff seeds have been locked in, teams are sitting all of their starters or some teams have nothing to play for.

These are all difficult scenarios for a set of mathematical models to be confident in. Last season, we identified and marked 17 of these games pre-emptively as "Stay Away". Doing this prevented the system from recommending either side of the game to users as well as preventing it from making any virtual wagers on those games.

A few examples:

Week 3 : CIN @ MIN - Both starting QBs Joe Burrow and QB JJ McCarthy were out, replaced by JJ Browning and Carson Wentz. More Vegas sportsbooks had the line at CIN +3, which given that the game was in Minnesota, is Vegas' way of saying "these teams are basically even, but we give the home field advantage 3 points". Minnesota won 48-10 and it was never close.

Week 10: JAX @ HOU - The Texans were without their starting QB and had to start Davis Mills vs. their division rival Jaguars who were a game ahead of the Texans in the standings at the time. The odds makers had this game at HOU +3 (which says a lot about their defense if they are starting a backup QB). Our system recommended the Texans moneyline bet, but we had it marked as a stay away so no bet or recommendation was made. Jacksonville was up 17-0 in the 2nd quarter and 29-10 in the 4th quarter before Davis Mills led a wild comeback, throwing 3 TD passes in the final 13 minutes of the game. So sometimes, the stay-away setting can cost you a win.

Week 18: LAC @ DEN - The Broncos had something to play for in this game (the number one playoff seed in the AFC) while the Chargers had nothing to play for as their playoff seed was determined. They decided to rest QB Justin Herbert and several key starters. Sure, the sportsbooks adjusted the lines to LAC +13.5, but are you ready to risk you hard earned cash on a team led by a backup QB and with nothing to gain by winning? Final score was DEN 19 LAC 3.

All in all, this cautious strategy resulted in flagging 17 games in the 2025 season as "stay away" games. The results of those games if they were wagered on would have been 4 wins, 12 losses and 1 push (thus avoiding 8 net losses). So in the face of volatility and lack of sufficient data, it's always best to hold off for a better opportunity. Remember, being profitable in sports betting is not only linked to finding the best opportunities, but also avoiding the uncertain ones.