THE BOOK--Playing The Percentages In Baseball

That graph.  It goes best season first, second best next, etc, correct?  I’m guessing that’s how it goes, but I wanted to clarify things.

Yes.  You’ll see that WAR is monotonically decreasing.  (And yes, at 60% of the reason I made this post was to use the word “monotonically”.)

It should be noted that both the average and replacement level lines were determined from players in the retrosheet era.  IIRC this only includes 19 total pitchers.

Jon (that’s my given name!), thanks for the word monotonic.  I learned something new.  I’m not sure why they do it that way instead of chronologically.  (Studes around?) When I first glanced at it, I was impressed by David Wells.  If you didn’t look at the legend, wouldn’t the red line appear to be the second best one on the chart?

Sorting by best to worst makes it clear who is the best, and where they rank.

As for the red line (Wells), he’s below the blue line (Mussina) for their 17 best years, in each and every year.  Wells is better in his extra years, but he is also in negative territory (worse than a replacement level pitcher) in some of those years as well.

The tough one is comparing Wells to Pettitte (if Pettitte retired after 2008).

#4.  The yearly were tried initially and it was got too confusing to compare the players.  Though the “monotonically decreasing” method is not perfect it makes for better comparisions.

Here is the url for the article where the Average and Replacement Level was set. 

http://www.beyondtheboxscore.com/2009/5/19/878150/war-graphs-average-and-replacement

The equations for the straight lines (hitting and pitching) is in the last comment.

THanks for that link, Jeff.  I figured you guys had an explanation somewhere.

Tango, eyeballing things, it looks like Pettite is a better WARrior.  I’m visualizing a trapezoid under his line that has more area than Wells’s positve triangle minus his negative one.

I like this format of graph, because you focus on career trajectory (even if it’s out of order) instead of total WAR.  David Wells is pretty underrated, I think, but has one of the flattest career graphs I’ve seen.  The lack of dominant seasons (no 5 WAR years) hurts him, implying he was never great.  However, after their best three seasons, he was better than Pettitte in every other season.

This type of thing also screams for a conversion of seasonal WAR to some sort of season HoF-Merit stat, something like WAR^2, WAR-above-4, (WAR-above-2)^2, etc.  Any ideas?  I’m not too familiar with attempts using pure value instead of “grey ink” or comparables.

Well, the answer is on Rally’s site, that shows that Pettitte with 45 WAR and Wells at 50 WAR (including negatives).  It’s close in any case.

The best match to HOM would likely be WAA (wins above average), excluding negative seasons.

I’m not trying to match BTF’s Hall of Merit, just looking for a merit(small m)-based system for measuring Hall worthiness.

Personally, I feel like each additional WAR in a single season is more impressive than the last, meaning 5 seasons of 2 WAR isn’t nearly as Hall Worth as 2 seasons of 5 WAR.  Same for WAA, but to a less extent.

How do you quantify those differences between each successive WAR in terms of “Hall worthiness” or “impressiveness”?

I think less is more. I would rather have outcry by people to get X player into the Hall than have people crying to get player Y out of the Hall. Electing Rice has now set the bar too low. It creates all kinds of problems. How many people would feel that David Wells is a Hall of Famer? Keep the bar set high and avoid the fight over borderline players.