So last Sunday we did grades for every NFL team through week 6. This Sunday, all 32 starting quarterbacks are rated using a new stat called QB Score, or QBS. What makes QBS different from QB Rating or QBR is that each quarterback’s rating is dependent on how their statistics compare to their fellow QBs.

This is another grab your coffee and donut kind of deep dive read, so if you do not have time for it, below you will find a chart with QBS, QB Rating & QBR scores and ranks for all 32 starters.

The average score for QBS is essentially zero (-0.05). Thus positive scores are QBs providing positive play, and negative scores show QBs providing less than average play. The further from zero, the better, or worse, the QB play has been.

If you like this kind of stuff you can learn all about QBS, how it is calculated, and why it may be a more useful way of measuring qb performance here: QB Score - Ratings all 32 NFL quarterbacks comparatively.

After week 9, I will re-run all the data and use it to hand out midterm QB grades for every team. The data below ranks QBs through week 7 stats, so a little preview of what to expect.

Since this is a brand new metric, it is open for modification to fine tune it. Please feel free to offer comments, suggestions or criticisms on this thread and they will be taken into account for the QBS calculations done post week 9.

Best of luck to your team today and enjoy the football.

​

QBS vs QB Rating vs QBR (Avg of QB Rating & QBR shown in last column)

  • defalt86@alien.topBEnglish
    1·
    11 months ago

    Any ranking that has Hurts at 10 for 2022 is flawed. Guy was untouchable. Top 3 minimum.

  • FloralAlyssa@alien.topBEnglish
    1·
    11 months ago

    Have you tested all the inputs for if the normal distribution is reasonable? I would think for some of them it would be, but for others you might have a longer tail on the up side (the elite QBs are better than the bad QBs are bad) which might explain why your final scores seem to be skewed.

    • Naskin@alien.topBEnglish
      1·
      11 months ago

      Yeah, each metric should be tested for normality. If non-normal, you need a fit that you can essentially correct things to being effectively normal (can use quantiles to create z-scores). Then each metric should be comparable. Once all are comparable, you can add a weighting factor depending on importance to generate your final QB score. Final score should have a normal distribution (can be converted at this point into some easier-to-digest number though, like quantiles, to give a percentage from 0-100)