Taking a longer glance at the WBC rosters

National team rosters for the 2023 version of the World Baseball Classic (WBC) were revealed on Thursday, and in many respects, they were more or less what you’d expect. A few teams look stacked, a few teams look like they’re just there to round out the field, and I’m sure some people gained knowledge they didn’t have about a given player’s country of origin and/or heritage.

The WBC remains a weird event in a weird metaphysical place. The intent is, as far as I understand it, for it to be the equivalent of the (FIFA) World Cup, but it’s definitely not the World Cup — this isn’t an article about the theoretical and real problems with the WBC, but that article almost writes itself and its contents loom over the entirety of the proceedings. The combination of key non-participating players (hey, where’s Aaron Judge? Where is nearly every elite MLB pitcher?), the fact that it takes place during Spring Training, and the reality that a tournament with knockout games makes little sense for baseball* make it a bit of a fool’s errand to analyze, but hey, I’m a fool engaging on an errand, here.

* In reality, knockout games only make little sense for baseball because in MLB (and probably other high-level pro leagues), the talent variation among teams is relatively low, so you need lots of games to reveal a “true” winner. For the WBC, the talent variation among some teams is actually pretty large, as you’ll see below — but the format is backwards in this regard. The knockout rounds happen after the weaker teams will likely be knocked out — they should’ve played knockout games at first and then series afterwards to make it slightly more sensible.

I started with the basic questions of, “Who’s got the best WBC roster?” and “How good is each team’s roster?” Answering this is actually really hard — for actual MLB seasons, despite there being more teams and way more games, and minor league systems and so on, there’s a better general sense of playing time assignments, and there aren’t stringent pitcher usage limits that interact in weird ways with the relatively short tournament in terms of whether pitching depth is actually important. The rest of this post is my very flawed attempt to answer the question. Here’s what I did. If you don’t care, and just want the rankings, scroll down to the table and team discussion.

Start with ZiPS

Yes, just ZiPS. Why ZiPS? Because ZiPS does not actually attempt to estimate playing time. Sure, some players will be projected for fewer PAs or innings, but that’s due to their own characteristics rather than playing time availability. In this case, this works in our favor: what we want to understand is something like “capability of production,” rather than “estimate of 2023 MLB production.” The problem with the latter is that it takes playing time availability into account, and given that many WBC participants are not projected to have an impact in MLB this year, it doesn’t make sense to use those projections for WBC performance, since they’ll be downscaled to zero for most players actually in the WBC.

So, in short: I downloaded every team’s roster, and slapped on the ZiPS projection for 2023 for every player on said roster. Easy enough.

Redefine replacement level

Okay, I’m not really doing this with as much rigor as it deserves. More like, “wave my hands in a vague fashion at the concept of replacement level.” In short, replacement level in MLB is really different from replacement level in the WBC. A few teams are basically MLB All-Star teams (well, only on the position player side), and quite a few teams have no MLBers or anything close. For the former, replacement level is way higher than your bog-standard 0 WAR MLBer. For the latter, replacement level is... I don’t know, the best player at your local D-III college? Maybe even worse? (That could be way off.)

My solution here is basically just to give extra “credit” to players actually in MLB systems, even if their ZiPS projections are kinda bad. This basically creates a separate, uniform replacement level that consists of “anyone participating in the WBC that ZiPS doesn’t bother to project,” (where “doesn’t bother” means “doesn’t appear on the list of 3,600 players with ZiPS projections for 2023 available on Fangraphs”), and scales everyone upward. My scaling is probably wrong, but I’m aiming for speed and generalities here, partly because I don’t think there’s actually a “right” way to scale it that would end up reasonably defensible in nature.

A real big honking note, here: this method fails miserably for countries with their own pro leagues that have relatively high talent levels — Japan (NPB), Korea (KB), and probably even Cuba (Serie Nacional). More on that later. Those teams remain wild cards.

Shrug at roster depth

So many problems rear their heads in trying to do this. Even beyond trying to figure who will/won’t start for each team, and how pitchers will be used, the damn rosters aren’t even all the same size. Great Britain (seriously, where’s Northern Ireland?) has 45 guys on their roster for some reason, while Canada and many of the teams from Asia have 30. The average roster is 35; the USA roster has 31 players while the Venezuelan roster has 38. And, I have no idea who’s going to play where for many teams, plus I’m guessing if any MLB-affiliated player even dreams about an injury, they’re getting recalled to regular Spring Training tout de suite.

Point is, as horrible as this is, my approach is basically average my replacement-level-adjusted-ZiPS-projected-WAR for each team. At least I split it by pitchers and hitters.

And then scale to 100

Because my replacement level scaling has no meaning, I scale every team’s average player quality to 100, where 100 is the best among all teams. I tried a few different ways of doing this, and the one I show looks meaner than the z-score-based one I also developed, but it doesn’t change the rank or anything.

Anyway, here’s the table, and then I’ll say some stuff about each of the tiers.

Thanks for coming

Australia, China, Czech Republic, Nicaragua, Taiwan

A quarter of the field basically looks like they’re here to be trampled. Unlike Japan and Korea, the China and Taiwan domestic leagues don’t really appear to be on a level that makes their lack of MLB-affiliated players a problem for my methodology. The best player in this tier is probably Taiwan’s Yu Chang, who has 0.6 career fWAR across 538 PAs, and was non-tendered this offseason after playing for four different MLB teams in 2022.

Surprising but unlikely

Great Britain, Netherlands, Panama

Not all surprising for the same reason. Did you really think that a team filled with a more than a few Curaçao greats would be in the same tier as... Great Britain? To be clear, the Netherlands has a better roster, because they have Xander Bogaerts... but beyond that, Curaçao isn’t adding that much. Great Britain has Trayce Thompson and a few prospect types; Netherlands has Xander Bogaerts, Andrelton Simmons, Jonathan Schoop, and fewer MLB-affiliated arms. The latter does have Kenley Jansen, I guess.

Panama has Christian Bethancourt and more prospect-type arms. Note that when I say “prospect” I don’t really mean real prospects, just like random arms that teams apparently didn’t care about enough to ban them from participating. Like the best arm in this tier might be Carlos Luna, a random non-prospect 26-year-old who struggled in Double-A last year. I don’t know why he has a better ZiPS projection than Kenley Jansen, that fact alone is deliriously funny. (Jansen has him beat on a rate basis for sure.) Also, where’s Johan Camargo?

Might make it interesting

Canada, Colombia, Israel, Italy

This is the tier where the rosters start to be populated mostly by major leaguers, or guys who have an okay chance of being major leaguers at some point. Fewer than 20 percent of the position players on these rosters lack a ZiPS projection; even slightly more than half of the pitchers have one.

Canada has Freddie Freeman, Tyler O’Neill, good prospect Bo Naylor, and not-as-good prospect Edouard “Vowels” Julien. Colombia has Gio Urshela and a bunch of role-player types. Israel has Dean Kremer, with one of the better pitching projections of the WBC. Joc Pederson and Alex Dickerson are the only outfielders listed for Israel, which is amusing. Italy has a bunch of major league position players, led by Vinnie Pasquantino, Nicky Lopez, and likely 2023 debutee/good prospect Sal Frelick.

Serious Rosters

Mexico, Puerto Rico, Venezuela

These rosters aren’t quite All-Star teams, they’re somewhere between that and “mostly fringe guys.” They’re basically like... major league teams.

Mexico has just one rostered position player without a ZiPS projection; Alejandro Kirk, Isaac Paredes, and Julio Urias are just some of their top-end talent. Puerto Rico seems like it should actually be stronger than it is — Francisco Lindor is a great starting point for any roster, and Javier Baez, Enrique Hernandez, Jose Miranda, are great complements, but it’s a little less deep than some of the other rosters. Venezuela’s roster is even better than these, featuring serious star power (Ronald Acuña Jr., Jose Altuve, Andres Gimenez) with a more-than-fine “rotation” core (German Marquez, Eduardo Rodriguez, Ranger Suarez). Venezuela has no non-major-leaguers on its position player side, and the guy with the worst projection for 2023 is Miguel Cabrera.

Heavyweights

Dominican Republic, USA

Essentially a bonafide MLB All-Star team on the position player side, the only downside to the D.R.’s WBC crew is that they don’t have much standout pitching behind Sandy Alcantara. They may just end up clobbering every other team into submission, though, because having Juan Soto, Julio Rodriguez, and Rafael Devers in the same lineup is terrifying for the relatively weak pitching slates every nation is sending to the tournament.

The only thing is... the U.S. roster is even more terrifying, and has somewhat deeper pitching. Anyone hurling to Team USA is going to have to get through Trea Turner, Nolan Arenado, a Will Smith/J.T. Realmuto combo, Mookie Betts, look how far down Mike Trout is on this list I’m typing, and also Kyle Tucker. There are 18 WBC participants projected for 4.5 or more WAR by ZiPS, 12 of which are either on the D.R. or USA team. As with most pitching staffs, USA doesn’t have a ton of top-level talent, but is fairly deep: Nestor Cortes, Clayton Kershaw, Lance Lynn, Brady Singer, and so on. Four of the 14 WBC-rostered pitchers with a ZiPS projection of 2.0 WAR or more are on Team USA, and as you slide down the rungs you basically have Team USA representing about 20-25 percent of the tournament’s pitching value while being just one team out of 20.

Wild Cards

Cuba, Japan, Korea

As said before, I don’t know what to do with these teams. It is worth noting that Japan’s roster has tons of potential — the few MLB-affiliated players on their roster are quite good, even beyond the main draw of Shohei Ohtani. If we had NPB projections that could be converted into MLB’s replacement level, I wouldn’t be surprised to see Japan in one of the top two tiers above.

Cuba and Korea are likely less impressive than Japan, but it’s hard to know where to slot them. Cuba has current MLB teammates Yoan Moncada and Luis Robert Jr., but unclear whether there’s much beyond that. Korea has Ha-Seong Kim and Tommy Edman, but are similar to Cuba beyond that, and the level of play in the KBO is thought to be less than that in NPB.

As always, my (messy, but “complete”) data are available on request.