Roster Construction: What's the best strategy to win in baseball?

[Vanimal46]

John LaRue with The Hardball Times wrote an interesting article titled "Evaluating Talent Distribution on Rosters"

 

He looks at teams over the last 30 years and compiled data to find out what is the best methodology to build a roster. Stars and scrubs, Deep with depth, Deep with stars, etc. 

 

Not every team has the resources and market size to build a super team. These methodologies help small market teams like the Twins compete against the Dodgers, Astros, Yankees, and Cubs of the world. 

[gunnarthor]

I'm not sure his methodology is correct but I've long thought it's best to have no sucky players on a team. Looking at jumps our Twins have made over certain years, a lot of improvement can be made by just getting rid of bad ballplayers and giving those at-bats or innings to better player. It happened in 2002, 2015 and 2017 (and I'm sure it happened a lot more than those years).  Having good depth helps you when injuries happen as well.

 

 

[Vanimal46]

 

I'm not sure his methodology is correct but I've long thought it's best to have no sucky players on a team. Looking at jumps our Twins have made over certain years, a lot of improvement can be made by just getting rid of bad ballplayers and giving those at-bats or innings to better player. It happened in 2002, 2015 and 2017 (and I'm sure it happened a lot more than those years).  Having good depth helps you when injuries happen as well.

 

Agreed. My takeaway when thinking about the Twins is they need a boat load of depth, and a quick trigger finger when a player starts failing. It's detrimental for the Twins when they give 300 some PA's to a replacement level guy like Danny Santana for example. Especially when they don't have the star power (yet) to make up for the bad play of others. 

[ashbury]

I have a related concern, not addressed by this study: loading up on modestly talented "hustle guys" who exploit the mistakes made by the other team, to compile so-so stats during the regular season. My speculation is that such position players may not provide much value at all if you succeed in reaching the post-season, where the opponents no longer make as many mistakes as average and don't just "give" you much of anything. Or, on the pitching side, guys with trick pitches (or whatever it is) that works well against AAAA hitters found in regular season lineups but are mysteriously absent come playoff time. Against them, you don't get 4 outs to work with on offense, nor do they let you off with 2-out innings when you pitch to them.

 

Every player likely suffers a decline due to these reasons in the post-season; but a guy like say George Springer starts with high enough value that he still contributes, while a guy with little regular-season value becomes a downright negative.

 

So, on that note, a big hole in this methodology IMO is that it doesn't have a way of addressing why a given "scrub" racked up less than 1.0 WAR; or, for that matter, *how* any particular WAR level was achieved (I haven't even touched on injury-shortened seasons). Of course I don't have any way in mind to help sort out the Nick Puntos and Jason Tyners and Scott Diamonds of the baseball world just from "WAR" types of stats. Scott Diamond gave us a 2-WAR season one year, but I wouldn't trust him in a playoff game any farther than I could throw him (which was about as far as he could throw his fastball).

 

Not being able to filter the data usefully doesn't mean the data doesn't contain the needed information. Anyway, this present study feels like it contains almost more noise than signal.

[Vanimal46]

Yeah even the author admitted to some flaws in his study. Justin Verlander for example was a 4+ WAR guy this year (a star in the study) but he accounted 3 WAR to the Tigers (depth) and 1+ WAR to Houston (depth)

 

You bring up good points, Ash. I just couldn't imagine the time needed to break down everything you've discussed in your post.

 

Overall I think it's good to see that a team with a lot of depth players can be more successful than a team of Stars and Scrubs. If we looked back, I'd be willing to bet the division champion seasons the Twins had were built the Stars and Scrubs way.

[nicksaviking]

 

Yeah even the author admitted to some flaws in his study. Justin Verlander for example was a 4+ WAR guy this year (a star in the study) but he accounted 3 WAR to the Tigers (depth) and 1+ WAR to Houston (depth)

You bring up good points, Ash. I just couldn't imagine the time needed to break down everything you've discussed in your post.

Overall I think it's good to see that a team with a lot of depth players can be more successful than a team of Stars and Scrubs. If we looked back, I'd be willing to bet the division champion seasons the Twins had were built the Stars and Scrubs way.

 

You better watch what you say before you utter something about Jason Tyner you can't take back.

 

I was encouraged by the quick trigger on organizational guys like Santana and Wheeler. I also liked the quick trigger on pickups like Wilk and Heston. Still, it would be nice if those kinds of guys only ever had a chance to pile up negative WAR only because of a rash of injuries and not actually be part of the planned depth.

 

[ashbury]

You bring up good points, Ash. I just couldn't imagine the time needed to break down everything you've discussed in your post.

It was exhausting for me just to write it.

 

And therefore I hope it was exhausting to read.