There are not usually major flaws in the top QB rankings, but I did sense the opposing defense can get overlooked.  In battles between top QBs and top defenses, rankers often keep “usually safe” QB choices at the top.  Is this catering to people’s hope for the obviously good QBs?  Was it possible that good streaming options got buried?

Verdict: Probably somewhat true, but still inconclusive.  I have shown my model can make marginal gains in overall accuracy, when compared to other reputable sources.  I would like to produce a more thorough analysis about which types of quarterbacks the model is better at projecting (overall top QBs compared to streaming candidates).