re: Most Overrated Statistic in Sports?

"Highest batting average by a designated hitter with the initials DH"


Heard this many years ago as Danny Hit stepped to the plate.

Guys, it is sacks and I'll explain why.

A. It is often an invisible stat. A sack can be negated by a penalty or first down. Getting a sack in either situation means nothing.

B. Sacks occur for a lot of reasons. For example, a player gets a coverage sack after the QB has been scrambling around for five seconds. Basically, he just made a glorified tackle, nothing more. Another example is when DC calls the perfect play and QB doesn't audible out of a blitz. Basically, a good percentage of sacks occur when the defender is unabated to the QB. An even more accurate example is when the QB actually does recognize the play, calls the right protection, but the RB misses his assignment, which happens a lot.

C. Then there is the fact that a lot of sacks come when the game is already decided. A team gets down by a few scores and starts throwing the ball a lot. Those sacks are like garbage points in an NBA game.

D. The worst part about the sack as a stat is it never tells a the whole story about a player. A player can get one or two sacks, but suck it up in pass coverage and run defense. He gets the headline but in the end the opposing QB throws for 300 yards because the same player that got two sacks can't cover the slot receiver.

It is almost a completely worthless stat. A better stat is QB hurries.

Hollinger's stats are all a joke. He's a hack, plain and simple.

This is going to sound weird but rebounding. Look at box scores, lot of teams that win get outrebounding

quote:

---

Nice work. The differences in calculating WAR between two such reputable sites is quite annoying. The only value of WAR is to get a general gauge of a player. Usually, a player with a WAR of 6 is going to be pretty fricking good, but other context matters so much more.

Telling me that Adam Jones has a 3.8 WAR (just made all the numbers up) is nice, but I would rather know he had 28 homers, 19 steals, and .294/.361/.497 line. A greater context is so much better than a fricking 3.8. What does that even mean?

---

Sports Illustrated had a nice article in this month's issue about WAR and its practical uses.

I wish each site would use their own version of WAR and call it something else. What is the point of calculating a stat with different formulas, yet calling it the same thing?

It's so hard to narrow it down to just one. Assists are absolutely absurd because it doesn't actually measure what the person did -- a perfect pass to a perfect shot doesn't automatically generate an assist because there's a chance the shooter misses it, just like a crap pass for a crap shot doesn't always not count because maybe the shooter makes it. Additionally, what happens when the individual makes the right pass and it causes a foul? It really is a bogus stat.

But then there's saves, which are equally crappy; other people have posted why. Even batting average has been marginalized by stats like batting average on balls in play, which is supposedly a better statistic.

I generally feel that the more statistics a stat draws from to generate a value, the more inaccurate it has the potential to be. Stats like PER, WAR, and whatever other you decide to use are only as good as the stats they are derived from and the formula from which they are derived. You use a crap stat to generate another number, chances are, it's a crap number.

ETA: For clarity, I'm not anti-sabermetric. I'm just anti-composite statistics. I think the statistical revolution is good, it just tries to generate a single number that says exactly how good a player is and you can't do that. To me, better statistical analysis breaks things down into more specific terms. To use assists as an example, a better stat would be the the number of times a player makes the "right" pass. Maybe that means the recipient of the pass has a good look at the basket, they're fouled, or it's a swing pass that then travels to the next person for a bucket. I know that's a HUGE can of worms because it asks questions about what is "right," but these are the type of things that teams look at when acquiring a player, I'm almost certain of it.

You can do the same thing with QBs, when the QB finds the open man, or a baseball player when they make a play on the ball and make a "good" throw. Stats like that compartmentalize a player's performance individually and prevents the complications of other screw ups. Flip side is how do you define "right" or "good."

quote:

---

This is going to sound weird but rebounding. Look at box scores, lot of teams that win get outrebounding

---

Most people don't understand how to measure rebounding just like they don't understand shooting efficiency

k's

Somehow if you strikeout 200 batters it's more impressive than a sub 3 era

Driving distance