1. We are generally of a mind that one should steer clear of what we term ‘incidental statistics,' such as:
"The last 6 times since 1978 that these two football teams have played outside at night, the road team won..."
"If a road team in NCAA men's basketball is coming off of 2 consecutive losses to underdogs, and the line is less than 10 points, the home favorite covers the spread 87% of the time..."
While this information makes for interesting conversation, and can sometimes appear to predict amazingly skewed probabilities, much of what would seem to be predictive is in fact quite random. As such, we find little value in these concepts. If one looks even casually at literally any data set over different time frames, it's easy to mistake coincidental randomness with what would appear to be patterned predictability.
2. We are also inclined to steer clear of progressive wagering systems (Kelly, Martingale, etc.). TO CLARIFY that statement, let us say that the logic of these can be interesting and useful in several applications, even outside of wagering. But to apply it to events whose outcomes are not purely random, about which one may also not have a solid probabilistic opinion, doesn't work for us. This is because sports games played by human beings are NOT the same as the turning of cards or rolling of dice, which possess no memory, biases, injuries, scheduling and matchup dynamics, etc. So IF one has a framework for making reasoned probabilistic judgments -- as LineAdvisor offers its subscribers -- we find it more logical to stick to similar wagers selected on probabilistic quality, than to compound one's risk by wagering on both the quality of one's ‘side' assessment, and toggling the amount to wager. Long ‘runs' will occur purely as a function of chance in almost anything, so toggling bet size can work so badly for a time that one is unable to recoup when chance swings the other way.