A Quick Attempted Measure of Team Depth

There’s nothing more enviable than a collection of stars. Generally speaking, people tend to focus on the top of a given roster, because that’s where you find the most impressive players, the players most likely to make a significant difference. The fantasy world operates like a video game with injuries turned off: you go in with your starting lineup and starting rotation, and you never need to replace anyone, and everyone’s competent. Every so often, teams do have sufficient health to be able to rely on the expected regulars. But more often than not, depth becomes a major factor. Sometimes this is even intentional — teams will accumulate depth in lieu of adding higher-quality starters. We’ve seen the A’s, for example, focus on depth, at practically every position. It’s not the most important thing; it’s just an important thing, frequently.

And it can be a tricky thing to measure, ahead of time. You can go about it by feel, on a case-by-case basis, but I thought I’d try something quick and easy. Which teams project to have the most and least depth for the season ahead, based on where things stand today? Presented below is not the final word, but it should be at least a starting point. Some numbers will change when, say, Max Scherzer and James Shields make up their minds, but they aren’t going to flip this graph on its head.

There’s no perfect way to do this, to my knowledge. This is just one way of doing this. For each team, I grabbed all the position players and all the starting pitchers who show up on the current depth charts. I decided to ignore bullpens, because they’re a lot more variable, and bullpen depth can also look a lot like rotation depth to some extent. So, one caveat here is that relievers are excluded, but individual relievers typically aren’t too important, and you can just mentally factor in that, say, the Royals have a more awesome bullpen than anyone else. It’s strong and deep. Anyway.

I then downloaded the Steamer600 projections for position players and pitchers. These are the regular Steamer projections, prorated out to full seasons so that every player is put over the same denominators. Then I simply counted how many position players or starters on each team are projected for a full-season WAR of at least 1.0. A threshold had to be set somewhere, and the threshold was going to be arbitrary no matter what, so 1.0 just felt right to me. These are players you can use without killing your team. At the minimum, they’re not quite good enough to start, but they’re better than replacement-level.

A graph of the results is embedded below. There are a few important things to acknowledge first. As noted, the cutoff is arbitrary — there are players who, for example, are projected for 0.9 WAR, and here they get no credit. This is inevitable, and it shouldn’t cause a bias, but it’s something. Also, this is a measure of the amount of depth, and not necessarily the *quality* of depth. The two are mostly related, but here a 2-win player is treated the same as a 1-win player. And then, obviously, this is just based on Steamer, and if you don’t believe Steamer for whatever reason, then you can argue with the numbers. This might be worth doing again when we have full ZiPS inputs, but the limitations of the projection systems means nothing is definitive. This is an attempt to describe a landscape. It’s as imperfect as you perceive it to be.

But you don’t want words anymore. You want a picture. Holy crap, Red Sox.

teamdepth1war2015

I forgot to note one thing earlier: I had to make a couple more judgment calls. Even though Steamer doesn’t have projections for Rusney Castillo or Yasmany Tomas, I determined, I think safely, they should both be at least 1-win players. So it’s all Steamer plus my opinion on Castillo and Tomas, and if you disagree with one or both, you can factor that in.

But that’s what I wound up with. The average team has about 15 players projected for 1+ WAR in a full season. Dividing that out, you’re looking at 10 – 11 such position players, and just under five such starting pitchers. The team with the most depth, by this method, is the Red Sox, with 24 total players. There’s a gap of four between the Red Sox and second place. Bringing up the rear: the Phillies, at seven, and then the Braves are at eight. The way the Red Sox break down, they have 17 qualifying position players, and seven qualifying starting pitchers. As far as the position players are concerned, that’s 17 out of 17, based on the depth chart.

So if you’ve been wondering why Steamer seems to like the Red Sox so much, depth is a huge reason. It’s not just the talent at the top of the roster. It’s that, when a starter isn’t playing, someone else pretty good should be playing. Most obviously, you can see this in the outfield, where Hanley Ramirez, Rusney Castillo, Shane Victorino, and Mookie Betts will combine at three positions, but based on Steamer you can almost construct a pair of complete lineups of 1+ WAR players. There’s just one player missing from the second team, and then consider that five teams have no more than eight position players overall projected for 1+ WAR. Some players on the Red Sox, surely, will under-achieve, but right now they seem well-equipped to deal with performance or injury adversity.

It’s not surprising to see the Pirates come out well, and this doesn’t even include Jung-Ho Kang. Sure enough, we find the A’s above average, with 14 qualifying position players and five qualifying starters. I was surprised to see the Twins show up so near to the front. What they lack in high-end talent, they attempt to make up for with higher-floor talent. For the record, included are Byron Buxton and Miguel Sano. The Twins could field a team mostly full of okay players. I’m pleased to have something nice to say about them. They seem like a team that could be able to successfully avoid the awful at the bottom of the roster.

Generally, I think the graph follows the order you’d expect in your head. Even if you quibble with individual numbers or projections, the trend appears mostly correct. Notable toward the back end: the Tigers and the White Sox. Neither is a shock, and in the latter case, this is a huge reason why Rick Hahn went out and brought in Emilio Bonifacio. He can complement a lot of players on that roster, and the White Sox need some insurance if they want to compete through adversity. Somewhat helping is that the Tigers also seem thin. That’s been a problem for some time, and these teams score less favorably than the Indians and Royals. The Royals also, as noted earlier, have that bullpen working in their favor, which doesn’t show up in the graph.

Real quick, I will take the chance to point out the Nationals. They’re a somewhat unusual team, in that they have a ton of talent, condensed into a few handfuls of positions. They’re overloaded with quality talent, and relatively weak in terms of secondary, supportive talent. One injury won’t be enough to sink the Nationals, but two or three at the same time would be pretty taxing, depending on duration. If the Mets or the Marlins want to have a chance at winning the division, they’ll almost certainly need for the Nationals to rack up a few DL stints, and Jayson Werth has gotten things off to an early start.

The Phillies and Braves are kind of depressing. The Reds, we’ve known about; they weren’t deep last year, either, and it cost them when so many players got hurt and/or under-performed. The Astros, interestingly, have 11 qualifying position players, and just two qualifying starters. Last year Brett Oberholtzer was solid, but he also had a 94 FIP- and a 112 xFIP-. You didn’t click on this post to read about Brett Oberholtzer or the 2015 Houston Astros.

So anyway, think of this as a starting point. As we get more information, we can update the numbers. If you have ideas, we can improve the methodology. And once the season gets going, depth will be differently important for different teams, and some lucky teams might have enough health that they never need to call on reinforcements. But you don’t want to get caught unprepared, since bad things almost always happen. As we stand, some teams will be crossing their fingers a little tighter than others. And the Red Sox? The Red Sox, if nothing else, look poised to make sure last year doesn’t happen again.



Print This Post



Jeff made Lookout Landing a thing, but he does not still write there about the Mariners. He does write here, sometimes about the Mariners, but usually not.


Sort by:   newest | oldest | most voted
jazzy jef
Guest
jazzy jef
1 year 4 months ago

Can you do a quick n dirty regression analysis for this vs. wins (projected or past)?

Darkstone42
Member
Darkstone42
1 year 4 months ago

I figured the Pirates would be near the top, given their depth of serviceable arms and actually pretty good bench this year. That’s a nice thing to have, as they learned last year, when the team survived two weeks of no Cutch.

Would you expect Kang to project at 1 win, though, to add to the Bucs’ apparent depth in this method? The two sides appear to have agreed to terms.

Rawson Baggs
Guest
Rawson Baggs
1 year 4 months ago

I recognize that any end-point would be arbitrary, but I’m just curious why you chose 1 win as the cutoff since the consensus seems to be that “major-league-average” is closer to 2 wins.

Darkstone42
Member
Darkstone42
1 year 4 months ago

I would argue in favor of Jeff’s 1-win cutoff since in order to contribute positively to “depth,” i.e. players who do not hurt the team when standing in for starters in limited action, either covering for minor injuries or to give them a rest, or who are utilized in largely favorable conditions (pinch hitting or getting a start against pitchers whose handedness favors his platoon splits, for example), do not need to be average. They just need to be above replacement level. Choosing one win, while arbitrary, is still definitively above replacement, but doesn’t put the requirement of “average” on the player.

Choosing two wins, on the other hand, would be useful in evaluating the depth of a team’s lineup and/or the quality of a team’s depth players, rather than the raw magnitude of depth.

Darkstone42
Member
Darkstone42
1 year 4 months ago

To go into more detail, evaluating depth would probably require some combination of a method like this (number of players who could be considered reasonably capable) coupled with an evaluation of the quality of that depth (perhaps paring out the starters versus the bench guys and seeing 1) how good the starters are and how deep the starting lineup is, and 2) how good the bench is), and then building some kind of “depth distribution” for each team.

But at some point, you’re just looking a whole rosters individually, and you don’t have this quick comparison Jeff seems to be going for, so I guess you have to find the balance between getting a reasonable approximation which fits inside your time and effort constraints (as well as your simple comparative visualization constraints) and getting the most full possible truth.

asdf
Guest
asdf
1 year 4 months ago

So what you’re saying is you’d like a page that takes projected performance, and weights it with educated guesses about playing time for each team at each position, and then adds those all up into team-based projections…?

….can i interest you in the “projected standings” page on this very site?

Nathaniel Dawson
Guest
Nathaniel Dawson
1 year 4 months ago

While “major league average” is considered to be 2 wins, the “average” major league player wouldn’t be expected to provide that over a full season. Pretty much all teams are going to be relying on quite a few players that don’t project to be major league average, and it’s important that they at least be not horrible.

Also, if you limit it to just 2+ WAR players, what you end up looking at is quality of starting lineup, since scant few teams have non-starting players that meet that threshold. There might be a half-dozen teams that have 2+ WAR players extending past their starting 8.

arc
Guest
arc
1 year 4 months ago

That’s an average major league starter.

Roger
Guest
Roger
1 year 4 months ago

I’m a little confused about the Mets placing low here because I just got through reading the article on the Mets’ ZiPS projections and that one seemed to make a case for the opposite, for them being a Red Sox-like team with few stars that could still count on average or better performance almost everywhere. Are ZiPS and Steamer forecasting the 2015 Mets differently for some reason?

SandyK
Member
SandyK
1 year 4 months ago

Five starting pitchers, seven position players (Cuddyer and Tejada just miss the cut) and Parnell, would be my guess.

Eric R
Guest
Eric R
1 year 4 months ago

If you looked at the two tables with zWAR in them in that post, you’d see a whopping 23 1.0+ win players… but those players weren’t adjusted for estimate play-time, so you have the pitchers combining for 3400 innings and the batters for 16000 PA.

The real numbers would be roughly 1450 IP and 6100 PA, so once you apply a realistic estimate for playtime, a load of those 1.0+ win players go away.

JK
Guest
JK
1 year 4 months ago

Actually, the number would would go up, as Jeff mentioned in his article he was looking at Steamer600. ZiPS has Cuddyer at being worth 0.9 WAR in 372 PA, so he misses the cutoff, but give him 600 PA and he’d make it.

However, that table also includes a bunch of guys who aren’t actually on the Mets (which is a pretty bad job by fangraphs, and it’s curious why no one has pointed it out).

Jeremy
Guest
Jeremy
1 year 4 months ago

ZiPS’ projections frequently have bench or minor-league players with their last-known-accurate team. This will often include unsigned free agents being listed with the last team they’d actually appeared in a game for. The top 90% or so of the lineups are usually just fine, but since we’re in the middle of the offseason when they’re released, there are often some stragglers (i.e. Jose Valverde).

Jim
Guest
Jim
1 year 4 months ago

I am fascinated to see how the Red Sox season plays out. It seems inevitable that they’re going to have starting caliber players on the bench, which looks ideal on paper, but I wonder if there are going to be veterans disgruntled with their playing time and if there’s going to be any challenges for Farrell to keep everyone happy

Rufus T. Firefly
Guest
Rufus T. Firefly
1 year 4 months ago

I’d be really, really surprised if they don’t make a deal that nets them another starting pitcher for a bat.

ozzie
Guest
ozzie
1 year 4 months ago

The White Sox are projected to have than twice as many players (and then some if you include the some members of the revamped bullpen) as they did last year (7).

It would be interesting to see graph of the difference in +1 players, between 2015 Steamer projections from current depth charts and 2014 totals, for all teams.

Noah Baron
Guest
Noah Baron
1 year 4 months ago

I’m so tired of reading FanGraphs articles that are solely based on Steamer Projections. Every time I see an article based on them I’ll just ignore it. They’re not accurate enough to be useful.

nerf
Member
nerf
1 year 4 months ago

Wait a minute – are you saying that Steamer isn’t a crystal ball with a view into the future?

Timothy
Guest
Timothy
1 year 4 months ago

You ignored it so well that you actually commented on it. Way to go!

Joshua_C
Guest
Joshua_C
1 year 4 months ago

Most empirical evidence I’ve seen suggests that Steamer is right up there with ZiPS in analyses of the most accurate projection systems.

I’m curious why you think they’re ‘not accurate enough to be useful,’ and whether you can suggest a more accurate means for projecting future performance.

CrazyPants
Guest
1 year 4 months ago

um, how about an informed opinion?

Joshua Choudhury
Member
Joshua Choudhury
1 year 4 months ago

Feel free to try and subjectively out-project Steamer.

Except, oh wait, FG already lets us try to do that, then publishes the results (along with ZiPS) to supplement Steamer. And the ‘informed opinions’ held by relatively knowledgeable baseball fans (i.e., us), consistently perform worse than Steamer and ZiPS. Even though we’re able to see the other projection systems when we make our own guesses, so we’re, theoretically at least, using them as a baseline and then interjecting our own opinions.

So maybe FG uses them because they’re the best publicly available way to estimate future performance. Which would seem to be a topic of principal interest to anyone looking to analyze the upcoming baseball season.

Nobody is suggesting projection systems are 100% accurate, but if you’re looking for an article where the entire discussion is based on the writer’s subjective, non-empirical opinion, then, I dunno, try BleacherReport? I bet they’ll have plenty of ‘Top 20 Bold Moves of the Offseason’ posts, and all of the articles will be empiricism-free.

Noah Baron
Member
Noah Baron
1 year 4 months ago

You want evidence that Steamer is terrible?

For pitchers, a single season of SIERA is better at predicting future ERA than Steamer projections, which take into account dozens of different variables and multiple years of data. And SIERA isn’t even that great of a predictor. I have my own proprietary metrics (structured around bbFIP) that predict ERA much better than SIERA.

Any so-called projection system that does worse than basic formulas that use one season of data aren’t good projection systems. This shouldn’t be controversial.

carpengui
Member
carpengui
1 year 4 months ago

Some sanity checks on their data would be welcomed. One of the Braves’ projections of 1.0 WAR+ (in other words, it’s counted on that chart) is for Rule 5 pitcher Daniel Winkler.

Not gonna happen.

Steamer projects 200 innings from a guy that had Tommy John surgery in August. Oops.

arc
Guest
arc
1 year 4 months ago

Looks like you need to work harder on not reading them, then.

Rufus T. Firefly
Guest
Rufus T. Firefly
1 year 4 months ago

I’m sick of reading all these Fangraphs articles about baseball. Talk about one-track minds!

Noah Baron
Member
Noah Baron
1 year 4 months ago

I enjoy FanGraphs articles frequently. I don’t enjoy articles that suspend logic and further reasoning because of Steamer.

B N
Guest
B N
1 year 4 months ago

We’ll take this under advisement and get our tarot card projections up by next week.

Dovif
Guest
Dovif
1 year 4 months ago

I like to use last years stats it is more accurate

Jay
Guest
Jay
1 year 4 months ago

I think depth is a very important, and frequently overlooked, consideration for evaluating teams.
It would be interesting to see how the graph changes after adjusting the threshold to 2.0 WAR and also 0.5 WAR.

Nick
Guest
Nick
1 year 4 months ago

The Nationals’ lack of depth is a real problem that has to be addressed when you consider that 4-5 of their position players (Zimmerman, Werth, Harper, Ramos, Rendon?) have significant injury history and Strasburg and Zimmermann are TJ survivors. Their combination of top-line talent, injury questions, and lack of depth make them a lot riskier than a lot of people talk about.

John C.
Guest
John C.
1 year 4 months ago

Well, considering that most of those guys with injury history actually got injured last year (Zimmerman only played 61 games; Ramos 88 and Harper 100, and both Ramos and Harper were clearly affected by injuries even when they playe) and the Nats still somehow managed to win 96 games while finishing 3rd in the NL in runs scored suggest that their lineup may not be as shallow as you think.

Pitching wise, starting pitching is the one area that they have some depth in Blake Treinen, Taylor Hill, Taylor Jordan and #3 prospect A.J. Cole. These guys are all #5 starter types, sure – thus moving their rotation from arguably the best in MLB (according to Buster Olney) to middle of the road. Combine that with the #3 run scoring lineup in the league (with all those injuries factored in) and you still win a bunch of games.

Bob
Guest
Bob
1 year 4 months ago

Rockies in the top third.

Can we get a good dose of health, please?

Fate
Guest
Fate
1 year 4 months ago

nah

Candy LaChance
Guest
Candy LaChance
1 year 4 months ago

Very interesting, Jeff! Not a big deal, but I think this system might be missing some starting pitching depth, based on how Steamer classifies pitchers. For instance, Steamer makes Zach McAllister (CLE) a reliever, even though he’s higher on the SP depth chart than someone like Cody Anderson.

Matt
Guest
Matt
1 year 4 months ago

2 thoughts:

1) It seems like maybe total WAR per team, according to Steamer600 (as opposed to normal steamer, as in the depth charts), might be more accurate. Then you would get a total sum that still includes 0.9 WAR players.

2) A maybe even better method would be to look at the Steamer projections, eliminate any player projected as a starter, and delete them, then sum the total remaining WAR.

Dave (UK)
Guest
Dave (UK)
1 year 4 months ago

I was coming down here to post exactly the same suggestion.

Owen
Guest
Owen
1 year 4 months ago

That might be interesting, but I don’t think either captures the general idea of what most people mean when they say “depth”. “Are the Tigers a deep team?” is a way of asking, “beyond their best 3-6 players, do the Tigers have good players to fill out the lineup, the bench, the backend of the rotation and the bullpen?” So “depth” DOES include starters… It asks whether your #8 hitter is a utility IF pressed into a starting role, or a potentially above average hitter; whether your fifth best starter is a washed-up journeyman or a proven young starter with TOR upside. It doesn’t take into account the performance peaks of your best players, so it doesn’t answer whether your slugging shortstop or dynamic center fielder will notch 3.5 WAR or 6 WAR.

In sum, total team WAR gives us too much information; team WAR minus starters gives us too little. Threshold cut offs are unfortunate, I agree, but that’s a problem with every system… The cutoff between “starter” and ” bench player” is just as unfortunate and arbitrary.

the inability to distinguish between 1War role players and 2.5 war core contributors is probably the biggest weakness of this system. It might be interesting to sum up the total war of the team outside the top 6-7 players… It answers the question, “Sure, they’ve got Strasbourg, Zimmerman, Fister, Harper, Rendon, Werth, and Desmond, but how deep are the Nats behind their core?” Again the cutoff is arbitrary but it gives a sense of overall depth.

The Juan Porciento
Guest
The Juan Porciento
1 year 4 months ago

If I were better at things, I’d color-code that graph to detail how many 2, 3, 4, etc. WAR players figure into each roster. A crude but eyeball-friendly way to integrate quality into the quantification.

But I’m not, so I’ll just fantasize about it until someone with skills pursues this line of inquiry.

jruby
Member
Member
jruby
1 year 4 months ago

I wonder how many of those Twins’ 20 are between 1 and 2…

Someone else! I call upon you to do the work for me, for I am at a bar watching Oregon forget how to tackle and Ohio State forget how to hold things without dropping them…

B N
Guest
B N
1 year 4 months ago

Funny. When I’m in the bar, it’s usually me forgetting how to hold things without dropping them…

CrazyPants
Guest
1 year 4 months ago

you said, “Holy crap, Red Sox” but I’d say “Holy crap, Twins.”

If this doesn’t show the weakness of this model, I don’t know what would.

Jason B
Guest
Jason B
1 year 4 months ago

You do know this isn’t supposed to be a definitive ranking of teams, right? This is just one prism through which team depth can be viewed. There are quite literally dozens of other approaches you can take, and you would have a slightly different takeaway from each.

arc
Guest
arc
1 year 4 months ago

You sure don’t.

witesoxfan
Guest
witesoxfan
1 year 4 months ago

Or maybe it’s showing that the Twins have been underrated and that they could be a decent, non-terrible team on the upswing.

siggian
Guest
siggian
1 year 4 months ago

The Twins look to be a team in which the backup players are decent enough and the starting players weak enough that it won’t matter that much in terms of who actually plays and what effect it has on the number of games won. That’s my take-away from this graph.

Agreed
Guest
Agreed
1 year 4 months ago

I had a similar takeaway.

Although it was interesting looking across: “Red Sox, Pirates, Twins, A’s, M’s, Yankees, Angels, Cubs, Dodgers, O’s… yeah I could see all of those teams being good this year. Wait a second!”

witesoxfan
Guest
witesoxfan
1 year 4 months ago

This is why I’ve been showing caution in the White Sox coming season. If everyone stays healthy and a couple of guys really step up (based on projections, 2B, 3B, and RF are the big areas where improvements can possibly be made with talent in house), they could have a really good team. However, it’s really only going to take one injury to really screw things up.

There may be other ways of determining some of the depth within each team, as suggested above, but represents a nice visual, if nothing else.

John
Guest
John
1 year 4 months ago

How about the average WAR for all the players divided by standard deviation? It takes into account power and depth. There are some flaws, but it is a simple and elegant estimator…

wpDiscuz