Originally posted by YACBros85:
Originally posted by 5_Golden_Rings:
Originally posted by YACBros85:
Originally posted by 5_Golden_Rings:
Originally posted by YACBros85:
Originally posted by SmokeyJoe:
Originally posted by YACBros85:
The most readily available statistics. Completion %, td to int ratio, ypa, total td's and total yards.
Ok. Rarely does one player sweep those categories so how are they weighted? For example, what if Trey Lance has more total TDs and yards, but his YPA, completion percentage, and TD/INT ratio are lower?
Its quite simple really. You rank them in each category. Then you add together their rankings in each category. Then you divide by the total number of categories.That should give you a total ranking.
In doing this you're essentially counting completion percentage twice.
Attempts = Yards/YPA, which means Completion % = Completions/(Yards/YPA)*100%. The bigger that denominator is, the lower your completion percentage; and the SMALLER that number is (Yards/YPA), the GREATER your completion percentage is. And how do you make that number (Yards/YPA) smaller? By making YPA bigger.
Thus, YPA already basically includes completion percentage, so there's no need to use it in a ratings system, unless you simply want to weigh completion percentage higher than other statistics, or you value it for some other reason than how it relates to you moving the ball down the field.
.
.
Another huge problem with this is that yards (and therefore YPA) doesn't distinguish the degree to which the QB contributes from the degree to which the WR contributes. Any QB rating system that doesn't use completed air yards instead of yards is already severely flawed, IMHO.
Weren't you the one a couple pages back saying you needed to tweak your formula more because Cousins being a top QB according to your formula didn't fit your narrative about him?
That was true regardless of how much I used Comp% (whether 0 or weighted up to ten times less than the other factors). My biggest issue was weighing TDs, INTs, and CAY.
It clearly had nothing to do with Comp% because Cousins had a worse Comp% than Jimmy and ranked significantly higher.
.
.
.
So, the issue is not related to completion percentage at all. It's related strictly to the respective importance of TDs, INTs, and CAY. To find that out, one will need to do some serious analytics, which I am presently too lazy to do.
Perhaps you aren't understanding me. You can use numbers to paint any picture you like. When you start by saying this formula doesn't work because the results don't fit my bias toward this play, this player and that player. It makes the numbers useless. You'll end up tweaking the formula so much that there will end up being very little difference between your results and your biased opinion. You'd waste less of your time just writing out a list of where you think each player should be ranked since that seems to be your end game anyway.
This feels like a slippery slope fallacy. When a result is clearly wrong, it makes sense to assume it's either an anomaly or the formula isn't as good as it could be.
We're talking about ONE player, Kirk Cousins, who clearly is not better than Aaron Rodgers. It is obvious that if your formula is ranking Cousins higher than Rodgers, your formula needs adjustment. What you cannot do, however, is do this for every result. It's only the
obviously wrong ones in which this is reasonable.
For example, if it ranked Jimmy as 11th, but you thought he was 18th, well, that 11th ranking isn't so unreasonable that it would be enough to throw out the formula. Why? Context and common sense in light of the total data pool, which suggests that the tier 2 QBs are closer to each other than they are to the tier 1 QBs. Therefore, it's completely reasonable for any one of them to have a wide range in which they would fall. But there is clearly absolutely nothing reasonable about Kirk Cousins ranking above Aaron Rodgers.
Breaking News!