Where I complain that winning the World Series in five games isn’t good enough:
In 2009 Mike Rizzo inherited the #1 and #10 overall draft picks along with the Nationals GM position, he was also fortunate enough to take over a team that would go on to receive the #1 overall pick in 2010 and #6 in 2011. These picks led to significant building blocks in the Nats three division wins over the subsequent five years. While there is no doubt about the success of these top picks, I began wondering where are the rest of these draft classes. Why aren’t there more Nats draft picks in the line-up and rotation?
So I started looking at the draft results for the first three years under Rizzo. Long enough ago so that the better players have started having some success at the major league level to be used to rate the draft. Player value was assigned to the team making the selection regardless of any subsequent trades. For this exercise I selected the Baseball Reference version of WAR as the point of comparison. In my opinion WAR is an overly convoluted stat but it is ideal in this case. WAR is applied to starters, relievers, and hitters so I could rate the picks on as close to equal footing as is possible, and just as importantly WAR is “stackable”, it can be summed up for multiple players per team to compare the totals. So even though WAR is not the perfect stat for comparing individual players it works great for ranking draft classes.
The big problem I had was that I needed WAR per player based on draft year and position. I checked several sites that listed WAR and could not find the draft info. So I did the work the long way. I used Wikipedia to get the draft order of players who made it to MLB and then checked Baseball Reference one player at a time. This introduced several flaws 1) unknown accuracy of Wikipedia, 2) data copy errors by me, and most significantly 3) players drafted by one team who didn’t sign and were later drafted and produced for another team. I tried to make sure that I only credited the WAR to the team where the player actual signed, and I removed the few duplicates, but I’m sure I have a few mistakes mixed in. Hopefully the volume of data is high enough to keep the margin of error low.
The Results – Great Drafts but no Diamonds
Once I had a table of the Team, Draft Position, and WAR per player drafted from 2009-11 I began running different sub-totals to see where the numbers led. The results were that the Nats had the third best total WAR, a very good result, although based on draft position they would have been expected to be #1. Still, nothing to complain about. The more interesting result was how skewed the Nats results were to the top 10 overall picks. Which makes sense, if you have the #1 overall pick you are going to spend the bulk of your time scouting top picks at the expense of looking for late round guys.
In the table below I have ranked each team based on different queries:
- All draft picks (Nats rank #3): This includes all draft picks during those three years regardless of draft position and WAR. The Nats have a total WAR of 59.6, with 55 coming from those four top 10 picks. A very good result considering that the average was 31.8.
- All “good” picks (#3): For this query I filtered out players with WAR <= 0. The idea being that even if a player makes it to the major leagues, if they are worse than a replacement player they aren’t worth counting. Not much difference in the results though.
- Top 10 Picks (#1): Looking at just top 10 picks eliminates a large number of teams who were fortunate enough to not have a top pick. The Nats towered over all other teams with a 55 WAR, far above the Os at 26.2 and Indians at 17.4. The Nats worst of four of these picks was Drew Storen, who has a career WAR of 5.2, exactly average for the top ten.
- After the top 10 (#28): This query was specifically selected to see how bad of a result could be attained by tweaking the numbers, I selected players with a >0 WAR chosen after pick #10. Not only because the Nats best players were in the top 10, but because the Nats didn’t have many draft picks in the first round after the top 10 (they did have two other round 1 picks that didn’t pan out). While this worked to get the desired result it really has no meaning.
- After the top 100 (#24): This is a much better representation of how the Nats compared with other teams for late round drafts, I picked 100 because it is a nice round number and because it is after the first three rounds plus comp picks (also filtering out players with negative WAR). Obviously 24th is not ideal, but that can be excused by the Nats focus on making sure that they didn’t whiff on those top picks. The Nats WAR was 5.9, the average was 14.4.
- Nice Picks (#28): This query was a way of looking for later round picks who were able to contribute to their team, selecting players drafted after #60 with a WAR >= 2. This query has the Nats looking terrible without me gaming the system. Only one player in three years, putting the Nats right near the bottom. Nate Karns was their one “nice pick” and with a WAR of 2 he barely qualified. Other teams had an average WAR of 17.1 in the category. (The Os and Padres were the only teams without a qualifying pick.)
- Diamonds in the Rough (zilch): These are great finds in the later rounds. I limited this list to players drafted after position 100. For the WAR filter I used a variable rate, 3.5 for 2009, 3 for 2010, and 3.5 for 2011. The Nats were one of seven teams to get blanked in this category. With teams averaging 11.3.
So the big takeaway from these results is that while the Nats took advantage of losing 298 games in three years by making very good selections with their top picks they missed an opportunity to add to that haul by not even making mediocre picks after the third round. The difference in 8.5 WAR between the Nats picks and the league average in the late rounds is likely a big reason why the team is in constant need of bench and bullpen depth.
I’d really like to analyze the draft results for the Bowden years and track how these results change as players add to their total WAR. But I’m really going to need a better way to collect the data, more accurate and less time-consuming.