Tips for getting started in KBO DFS

It’s seven o’clock on Saturday morning and I am sipping my coffee watching a baseball game being played on the other side of the world. I have a substantial amount of money on the over in the Korean Baseball Organization gake between the NC Dinos and the Doosan Bears. It’s 3-3 in the top of the eighth and the O/U is 10 runs; starting pitchers Sung Young Choi vs. Raúl Alcántara.

Don’t get me wrong, I was sweating the bet pretty good at this point, but I still felt confident now that Doosan had gotten into NC’s bullpen more runs would be pushed across. The Dinos failed to score in the top of the eighth and I was sitting on 1.5 innings to get 6 runs… then Doosan exploded for nine runs in the bottom of the eighth off Jae-Hwan Bae and Young-Gyu Kim. It was the start of a good weekend.

KBO has definitely filled the void of no sports for those of us grinding daily at 5 a.m. The league has given us a ton of drama and Draft Kings introducing KBO DFS has really added a fun way to build bankroll as we anxiously await the return of American sports.

One of the many edges in KBO DFS has been the ability to get your hands on advanced data and metrics. Available data was few and far between (especially in English) for the first two months. Currently, FanGraphs has individual season stats for pitcher and hitters. Baseball-Reference and mykbostats.com have added some much-needed team and park stats for those of us creating projections. And Pitcher List did their thing covering pitcher performances and providing great daily/weekly summaries. However, there really wasn’t any team bullpen-exclusive stats available anywhere and we know how much of an edge bullpen stats can provide when trying to correctly predict player outcomes and beating Vegas lines.

It took a little bit of time, but I compiled individual stats for all 214 KBO players who have seen the mound for at least one pitch this season. Then I took out all starting pitchers and used some discretion when including a bullpen pitcher making a spot start but lasting less than a few innings. I then calculated the stats below per team. It’s not perfect, but should give us plenty of insight into the mess that is KBO bullpens.

Updated as of 6.29.20

Ok, let’s dig in!

1. Samsung and Kiwoom are the only decent bullpens, still far from great but standing out from the pack at this point in the season. Kiwoom is anchored by Sang-Woo Cho, who is 11 for 11 in save opportunities and also racked up 2 wins from the bullpen. He has an 0.52 ERA through 17.1 IP and a 25% strikeout rate. He is one of the few consistent relievers in the league and should support Kiwoom’s starting rotation, locking down the +4 DK wins points. Including the starting pitchers, Kiwoom has a 4.14 ERA, so their 4.40 bullpen ERA is close to that of their starters. Samsung only has 10 save opportunities so far, but is 9-of-10 as a unit (7 coming from Kyumin Woo). They have the second-lowest bullpen WHIP and are allowing far less than the league average in DK PAPI.
2. “xRPI%” What? Yes, it is a stat I came up with after getting frustrated that I couldn’t find inherited runners allowed percentage or anyway to manually calculate it. IRA% is a key metric I use when analyzing MLB bullpens. Does the team have a staff that can come in with two outs and a runner on second and get the final out and preserve the win for my starter? That is many times the difference between cashing a double up. xRPI% is my attempt to figure out how many runs scored by the opposing team are inherited or directly from the bullpen. It is not perfect as it doesn't factor in unearned runs from exclusively a relief pitcher (errors, etc.) and it doesn’t tell me exactly how many inherited runners didn’t score, but it will have to do for now. The cool thing about xRPI% is that it can be used two ways — as a predictor for SP W% and thinking about stacking hitters against a bullpen. For example, the Lotte Giants have a 1.23% xRPI; meaning almost all of the runs given up by the bullpen have been earned by them and not inherited. While this typically bodes well for a SP who gets pulled in the middle of the inning, it also means that the Giants bullpen is giving up about half a run per inning and there is no blame on the SPs for that.
3. The best team in the league by most metrics, the NC Dinos, arguably have the worst bullpen. When I first started digging into this I almost didn’t believe it and assumed there was an error in my spreadsheet with a formula somewhere, but after looking at box scores and sweating (or not sweating) some Dinos Overs, it started to come together. A league-worst 6.55 bullpen ERA (compared to a team total 4.71 ERA), .77 RPI, 65% save percentage and a below-league average BR% all spell for some long, late innings for the Dinos. The good news is they have the best pitcher in the KBO, Chang-mo Koo (my man crush), and an explosive offense, so even when the bullpen gives up 3-4 runs in the eighth and ninth innings, the Dinos are many times still in a position for the victory. All that said, the Dinos do have two arms in their bullpen that can shut down the opposing batters: Jong-Hyun Won (2.21 ERA, 0.89 WHIP, and 12 saves) and Yoon-goo Kang (2.61 ERA and 1.84 WHIP). The health of these two and a need for a third consistent reliever will be key if the Dinos want to continue their run and take home the championship.

This data is not easily accessible as much of it was compiled manually, so use it your advantage daily. Whether you like to look for edges in team totals and game totals or are deciding between two stacks for DFS, factor in the bullpen metrics! How long does the projected starting pitcher last on average? Is he under any kind of pitch count? Which bullpen pitchers already pitched multiple times this week? Who is likely to pitch in a long relief spot? What is the Vegas F5 run line? Predicting how many innings a team will face a bullpen and who is available is a huge edge that we want to hammer before this data becomes more readily available.