Whew...it's a good thing you have all those strawmen to carry the goalposts around the field for you.
1) GF% was the biggest stat last year! (Not really. It wasn't cited in isolation. No one's saying this.)
2) Stanley has the 2nd best GF% on the Jets! (Not meaningful in small sample, his shot metrics and modeled stats range from team worst to meh).
3) GF% wasn't luck driven last year! (Wasn't used in isolation, much larger sample. No one's saying this).
4) If Stanley was bad the numbers should say that but: 2nd best GF%! (Most of the numbers do say Stanley is bad - except GF%, which we know is highly susceptible to luck).
5) To raideralex99 - My problem with the stats nerds is that just because Stanley has 2nd best GF% doesn't mean he's the best defenseman. (Again, no one is saying this besides you.)
6) "According to your claim (before you started acting like you've dunked on me), Stanley should be giving up the most xGA/60 on the Jets. Turns out you'd be wrong" (Please show me where I made that claim. Once again, no one's claiming this. You pedantically latched on to me saying "giving up more" instead of "has a worse ratio of". In net terms does it make a difference?)
oh good lord there's more!
"Stan is firmly in the middle of the pack for Jets D men, doing a good job defending vs the players he's put out there against"
- Oops! Your bias is showing - you've got that table sorted in reverse, but whatever. He's 5/7 on the Jets in xGA/60. I guess that settles it.
"So when you say that he's bad at defending, that's statistically false. He actually needs to improve on generating shot attempts for both him and his teammates"
- Again, I never said he was bad at defending, although I will say he's not particularly good at it by any measure.
"And edit: the CSV sample size that you called "big" was 192 minutes last year. Stans sample size this year is 530 minutes, and you called it "small""
- I never said the CSV sample size was large, I said the C+S and E+S samples were much larger than the Miller+Stanley sample. Also, the CSV sample was bad across the board, not just GF% (and GF%, as I may have mentioned, is not particularly useful in small samples in isolation).
All I've been saying, over and over, is that GF% is a poor metric to use in a small sample and in isolation to judge Stanley by. Luckily, no one is using it that way except you! Every other metric/analytic has him ranked low...except the GF% outlier, which we already know is highly susceptible to luck.
So in summary, there's no "one stat to rule them all". GF% is highly susceptible to randomness. Larger samples are better than smaller samples. Miller is better than Stanley (but if you have to have Stanley out there every goddamn night, putting him with Miller looks like the best bet). It would be nice to see if anyone else is good with Miller but Stanley Über Alles...