Visual Baseball:  Ranking the AL EAST Starting Rotations

Hi there. I’ve been experimenting with ways to visualize starting rotations using the Rankometer format. Here are two approaches I’m playing with:

APPROACH #1 ranks starters by ERA and puts these rankings into 5 tiers (each tier consisting of 14 pitchers). The first tier consists of the 14 best starters in the AL, and a pitcher in this tier could be described as “pitching like a #1 starter.” Pitchers in tier 2 (the next best 14 starters in the AL) could be described as “pitching like #2 starters,” and so on. So if you’re a Yankee fan you could look at your rotation with this version of the Rankometer and say “Right now we’ve got three #1 starters, a #2 starter, and a #5 starter.” So this is a simple, at a glance way to sum up a rotation.

APPROACH #2 is based on the idea of rotations being matched up head-to-head. In a theoretical 5-game series between two teams, each team would trot out their best starters for Game 1, their second best starters for Game 2, and so on. Using this philosophy, in column one (entitled BEST STARTER) we list each team’s best starter, ranked by ERA. In column 2 we list each team’s second best starter, and so on. So if you’re a Red Sox fan, your takeaway is “Right now our best starter is about average, but our #2,3,4 and 5 starters stink relative to the rest of the AL.” So this flavor of Rankometer allows you to sum up a rotation in a slightly different way. It also uses the graphic equalizer metaphor to communicate a rotation’s strength.

Each approach communicates slightly different things and has its advantages & disadvantages. Which do you like better?

APPROACH #1: Starter Rankings

image image image image image

APPROACH #2: Rotation Rankings

image image image image image


Kevin Dame is a writer and visual designer who brings sports information to life in new and meaningful ways. Visit his website and follow him on Twitter @kevintdame.
15 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
bob b.
13 years ago

my two cents: although the second method seems more visually impressive to me, i have to say that overall i prefer the first method as it seems to offer more of a complete picture (of ordinal ranking).

Ahmet Hamdi
13 years ago

I like method #2. Good graphic method of having a quick take away of the pitching strength of a division/league. Granted, I would have two recommendations.

1) ERA deals with past performance based on runs scored against, and in such a small sample size of games, ‘luck’ can cause a sizable change in ERA. Try using a pitcher rating that focuses on evaluating the players in a more future focused manner?

2) I feel that this could be used on fielders pretty well. You could even use a baseball diamond background and use a color scale of blue -> red in each fielding ‘zone’ to rank a team against the field.

Sam
13 years ago

The second method seems more appropriate for looking at playoff matchups, since a) in the playoffs, teams will tend to put their best-performing starter at #1 (rather than their “#1” starter at #1) and b) the matchups will stay on track (as opposed to the regular season, where #1 starters don’t always face other #1 starters.

Ahmet Hamdi
13 years ago

That is very true. However, having a division/league relative ranking can give novices a quick look at seeing ‘is our rotation after pitcher #3 a weakness?’ ‘should the team bring up a minor league player that projects to perform better than your current #4?’ (I refuse to believe in a #5 quality pitcher)

Absolutely, Method 1 is best at giving an absolute look across a division/league where the wealth/quality is being accumulated, but Method 2 helps tell the story of how well distributed it is. Now does the distribution of ‘wealth’ across a pitching rotation matter? Well, I guess you could have asked Z. Greinke last year how if it would have helped the team or not …

Ryan
13 years ago

I like it, but agree with Ahmet that you should a different metric than ERA.  I would probably go with WAR, but if you feel tied to ERA, you could at least use ERA+, or FIP or xFIP.

Chris
13 years ago

I appreciate the second method slightly more simply because it gives you a sense of pitching depth. The first method seems to emphasize the front of the rotation.

jfpbookworm
13 years ago

I might actually go with something more color-based.  Set up a gradient where the best starter is in red, the best “tier 2” starter is in orange, the best “tier 3” starter is in yellow, the best “tier 4” starter is in green, the best “tier 5” starter is in blue, and the worst starter is in violet, then use gradients to assign intermediate colors to the starters in between.  Then, when you put the starters for a team side by side, you get a spectrum where reddish = good, bluish = bad.

kardo
13 years ago

I like the second method. The first method doesn’t add visual info over regular sorted pitcher lists.

KenZ
13 years ago

Method 2 gives a clearer picture of how your rotation lays out.  I think you could enhance the look by giving the relative value of Starter 1 by spreading them on a linear scale of 0 – 7 ERA (above 7 is worthless) with 0 at the top.  So if the ERA spread between adjacent pitchers is more visible. For example, if Petite has a 3.10 ERA and Fister is 3.97) – there ranked 4 & 5th for #1 but in no way that close.  Conversely, we could also see if the difference between being 2nd and 6th might be negligible.

Baseball DVDs
13 years ago

your site impressed me a lot you are doing nice job that will me and all the other in future keep posting this type of stuff.

Geronimo J
13 years ago

I like both sets of stats.  The first provides an overview of the total strength of any teams rotation, and the second allows for for an understanding of where a rotation is strong and where it’s weak.  I’d like to see both of these stats available and updated daily for both leagues.

jimmy james
13 years ago

This is great, great stuff.  I’d love to see these for the rest of the league if you have the time.

GreggB
13 years ago

I think both do equally well at presenting the strength of a team’s starters.  But #1 does a great job at showing the relative value of individual pitchers at the same time, while #2 distorts that data.  How good has AJ Burnett been?  Using #2, you only know that he has been great for a #3 starter.  But #1 shows he has been the 14th best in the league.  That is real additional information about individual players, delivered without compromising data on team performance.

Grouping by best, second best, etc. is creating false competition; this isn’t the way the game is played.

Finally, #2 creates a situation where an individual pitcher will bounce around from category to category based on how his teammates do. You could be on top of the fourth best pitcher tier one day, then suddenly relocated to the middle of the third best column the next day—even though you didn’t pitch (because your teammate had a bad start).  Too erratic.

GreggB
13 years ago

First of all this chart is definitely valuable.  Offering some basic splits would make it even more so.  Right now it is dealing with a very small sample size.  It would be great to click a button and see 2009 ERA, last ten starts (spanning seasons)ERA, career ERA, FIP, etc.