Stealofhome's Fictional Feeder Leagues and MLB Extravaganza

[stealofhome]

AL Tomas Reed MVP: Nick Reed, CF, Cleveland Indians - 7.9 WAR, 142 wRC+, 10.5 ZR


AL Top Pitcher: E.J. LaFleur, RHP, Detroit Tigers - 7.7 WAR, 70 FIP-, 1.9 K/BB


NL Tomas Reed MVP: Flavio Viera, 3B, Florida Marlins - 7.6 WAR, 160 wRC+, 1.2 ZR


NL Top Pitcher: Bond Lott, LHP, New York Mets - 6.2 WAR, 81 FIP-, 1.8 K/BB

AL Rookie of the Year: Diego Silva, CF, Oakland Athletics - 4.5 WAR, 112 wRC+, 12.1 ZR
NL Rookie of the Year: Robert Chisenhall, SS, Cincinnati Reds - 6.5 WAR, 115 wRC+, 20.6 ZR

[stealofhome]

Nathanael Clark, 1B

Clark was a lefty-batting slugger who was drafted 3rd overall out of UC Santa Barbara in 1929. He went to four all-star games and won two silver sluggers. He also added a gold glove to his trophy case in 1936. He retires as possibly the 2nd or 3rd best true 1B behind Aiden Massicotte and Elliot Raprap.

Career Stats: 9061 PA, 2129 H, 186 HR, .361 wOBA, 121 wRC+, 82.4 ZR, 46.2 WAR, 37.9 JAWS

Martin Knisley, RHP

Knisley's entire career was spent with the Kansas City Royals, where he went to one all-star game and won a World Series. He was a 7th overall pick out of Florida State and finished in the top 3 of the Pitcher of the Year voting twice (1933,1941).

Career Stats: 3505.3 IP, 1133 BB, 1238 K, 3.48 FIP, 90 FIP-, 64.3 WAR, 46.7 JAWS

Victor Wade, RHP

Wade split time primarily between the Tigers and A's but also spent his last two years in the majors with the Indians, where he won a World Series. He won one gold glove and went to an all-star game. Not too bad a career for a 3rd round draft pick out of Rye High School.

Career Stats: 3108.3 IP, 945 BB, 1025 K, 3.85 FIP, 98 FIP-, 44.0 WAR, 45.2 JAWS

Corbin Cuva, LHP

Cuva spent his whole career as a reliever for the Brewers - much of it as their primary closer. He went to four all-star games and was the Reliever of the Year twice.

Career Stats: 866 IP, 307 BB, 434 K, 302 SV, 3.10 FIP, 83 FIP-, 21.9 WPA, 15.4 WAR, 24.2 JAWS

[stealofhome]

We're about to enter the third era of Major League Baseball. Although there is no color barrier in this league, I am expecting the statistical outputs to change significantly. I'll be using this as another opportunity to look at some various things.

First is updating the minor leagues to match the current real-life MLB.

This is what I have and I've pulled what I could from the 2019 updates thread.

[stealofhome]

I wanted to do a complete game sanity check. It looks like the OOTP league is actually less extreme than in real life in terms of complete games per games started by pitchers. This article on Baseball Prospectus is a pretty good history of pitching usage.

[stealofhome]

I've attempted to replicate Stephen Loftus's work here with my league. This primarily involves using a weighted average of PA or IP to determine what a "Season" means for a player. Then take the average WAR/Season based on the draft pick and origin. It appears that college arms are the most sure bet and that there is a steady trend up through the 10th Round. From there, small sample size and random changes in talent level really make things get a little crazy. This only accounts for players who were drafted and made the Major Leagues, so only a very small percentage of players past Round 10 show up here.

[stealofhome]

The next thing to check was how players stacked up in terms of WAR per season based on their origin. Again, this doesn't take into account the fact that there may be a smaller likelihood of the player actually getting to the major leagues based on where they started. Since players that don't make the majors are deleted, I don't know how to work that into the analysis.

[stealofhome]

I did a little bit of a look at catcher defense, though it is so hard to determine much with a multitude of variables. I limited this to catchers with at least 100 innings played in 1946. I then converted their runners thrown out and stolen bases allowed into a run value, based on a stolen base = 0.2 runs lost and a caught stealing = 0.4 runs added.

The other thought is that runners will be less aggressive with a catcher that has a better arm and those are the other charts. There does appear to be a little bit of an uptick with some of the extremely high rated arms, but that is a very small sample in the minor leagues.

Max Weinstein has an old post about the overrated value of the catcher's arm. It does appear to have a significant impact in OOTP, but in real life he found that most stolen bases happened due to the pitcher.

[stealofhome]

Another look at catcher ERA - in all leagues and just in MLB. There again appears to be an uptick in the 130-150 catcher ability range with a strong downturn past 150/200. There aren't enough good stats connected to the catcher to attempt a better analysis (strikeouts/walks). Sansterre ran a test on how catcher ability impacts walks/strikeouts.

[stealofhome]

I'm trying to run an analysis on the batter/pitcher matchup but running into a few issues. First - in the players_at_bat_batting_stats export file, the "result" field is coded with a number. I was able to figure out was most of them meant, but there are no events of "3" in my file and very few "11." I have no idea what 3 may be or if it isn't coded to anything, but I believe the 11 is supposed to be a fielder's choice. However, there are many instances where a fielder's choice is not added to the database, so it looks as though it never happened.

Also, I wanted to create a league-specific run expectancy matrix, but don't think it's possible from what I have. The "run" variable in the at bat export file doesn't signify that a run was scored during the event but that the player eventually scored while on base. I figured this out while looking at why more runs were scored on walks than singles. I may just have to use Tom Tango's or something even though it's not relevant to this fictional league. Both missing and wrong data are throwing me off here.

Edit: I did just notice that the rbi column is what I was looking for, though it still is not perfect. There is also the possibility of using lineup spot to make the at_bat data sequential and building a run expectancy from that.

My eventual goal is to analyze which ratings on both the hitting and pitching side impact the matchup the most and lead to better results either way.

[stealofhome]

Okay I'm pretty pleased with this one - the first step to the hitter/pitcher interaction is to create a run expectancy table that is specific to my league. This requires play by play data (Very High Stats detail). I found an awesome article that walked me through the second half of this, but I still had to get the data in the right shape from OOTP.

I first had to order the data by event in the game instead of player_id and then calculate the number of runs after the event.

This table calculates the average run expectancy for each baserunner/out state in 1946 in my league. I'd say for the most part it lines up well with current MLB standards.



The other two charts are the distributions of expected runs at each game state and the mean posterior estimates of observed samples in 1946.

I can give the code to anyone if they're interested. I may just start a GitHub page for all of this.

[stealofhome]

Once the expected run value at each event was in place, I just needed to compare the run value before and after each and group for the average. Here are the custom linear weights for my league in 1946 which can be used to calculate wOBA, etc.

[stealofhome]

The Extravaganza has moved to OOTP20.