Investing & Maybe Fantasy Baseball

I play fantasy baseball at NFBC. Specifically, 12-team mixed leagues. My analysis and numbers are geared towards NFBC’s format with two catchers in your starting lineup.

I’ve noticed there’s still chatter about position scarcity being a real thing still other than catcher. I agree position scarcity does exist at catcher, but at other positions is just utter nonsense.

A starting lineup consists of 14 hitters for a total of 168 hitters. During the draft your goal is acquire as many of those 168 for as cheap as possible. If you draft the 200^th best hitter at pick 150 you’re losing value. (I’m assuming global values with that comment; if you believe he’s the 150^th best hitter available then you did not overpay.)

Below are the bottom three non-catcher hitters. It just so happens they fill each major category: Outfield, Middle Infield and Corner Infield.

I rank players via the SGP model. Basically, it gives a value for every hit, run, home run, rbi and stolen base. The higher the SGP, the more valuable a hitter is.

According to my projections there isn’t much difference between the three hitters. Michael Chavis is 0.005 better than Andrelton Simmons, which is virtually nothing.

You may be asking, “but Matt what about second base or third base?”. Good question. At the end of drafts you’re usually filling your middle infield or corner infield spot so it doesn’t matter what position they play if they’re filling those slots. Obviously, if you have drafted a third basemen at the end of the draft then you’ll have to A) take someone out of the 168 best hitters or B) plan to take a third basemen earlier in drafts.

The next logical question is are certain positions deeper than others? Yes. Third base is deeper than first base. Shortstop is more top heavy than second base, but second base has more value in the middle. Shortstop, at the bottom is thinner than second base. Even though you have to start five, outfield is the deepest position this year.

All that said, if you’re picking in the fifth round and you’re debating between Matt Olson or Whit Merrifield, take the best player available.

I was listening to a fantasy baseball podcast that have “expert(s)” as hosts. I’ve been listening to the show for a while and I never understand their rationale for how they evaluate players. (After I’m done writing this I will stop listening to the podcast.)

Humans have a difficult time hearing ideas that are different than our beliefs. If I think Jose Altuve is overvalued I will only like to hear from people who agree with me. Or if I hear someone discuss the pros and cons, I will only remember the cons part of the argument. To avoid this bias, I do my best to listen to ideas that contradict my own; in fact, I try to uphold this belief in all aspects of my life.

So, I don’t mind if people have differing opinions than me. That said, if someone has an opinion, I want evidence that backs up their viewpoint.

Below are the 2019 statistics from FanGraphs. The ADP data is from NFBC (date range: 1/25 to 2/6). Of these two players which would prefer?

Player A is Tommy Edman and Player B is Jeff McNeil. In the podcast the hosts swooned over McNeil and shitted on Edman. I don’t understand the difference in ADP and/or why the podcast hosts felt the way they did. (If you listen to the podcast they do provide their reasons but it’s mostly anecdotal and chopped full of personal biases.)

In 1974 Paul Slovic and Daniel Kahneman did a study with professional horse handicappers. There were five rounds of betting. Each round consisted of ten races and with each subsequent round they received information of their own choosing. So, in round one they were given five pieces of information. In round two ten pieces and all the way up to 40 pieces of information.

What the researchers found was their accuracy didn’t improve with more information. However, what did improve was their confidence level. That’s pretty striking. It shows that after a certain amount of information, any extra information contributes to our earlier conclusion (i.e. confirmation bias). Also, if the new information disagrees with our conclusion we ignore the information.

Obviously both players had limited sample sizes last year (more so with Edman) but I used the numbers in the table above to come up with my projections for both players. In my experience the numbers I cited above are all the numbers I need to make a projection on a player; any more data or observations contributes to bias.

Among second basemen Edman and McNeil are ranked 9^th and 11^th respectively. I think these projections are reasonable. Edman’s numbers last year suggest he wasn’t overtly lucky, that he should be able to at least a be productive player for the Cardinals.

Edman’s detractors say he doesn’t have an everyday job and will begin the season as a super utility player. RosterResource has Edman as its everyday left fielder.

If you don’t believe Edman will play every day then who is going to play left field? Dylan Carlson is probably the best bet, but he has only played 18 games in triple-a. Tyler O’Neill? He’s not a good defender and I doubt he hits above .240 if he played every day. For his career he has 58.2% contact rate. To put that into perspective, Chris Davis in 2019 had a 62.9% contact rate, the lowest among all hitters with at least 300 plate appearances. Lane Thomas? I don’t much about him, but his scouting report suggests he’s maybe a 4^th outfielder type. Also, Dexter Fowler and Matt Carpenter only have jobs because of their contract, Harrison Bader, who they demoted after he got off to a slow start. In my mind there is far greater chance Edman gets 540 at-bats than 350-400.

I don’t want to make it sound like I’m being petty and/or overly critical about the hosts of the podcast, but there have been further evidence that I don’t believe they have any idea on how to evaluate players.

Another example is Matt Olson. He’s currently being drafted as the 4^th first baseman, 58.84 ADP. The podcast hosts had no idea why Olson was being taken so high. They said that he would have to hit 50 home runs in order to be taken that high.

Let’s look at his numbers. The table below shows the statistics for his 2019 season, his prorated 2019 season (if he played a full year) and my 2020 projection.

Olson is currently my 4^th rated first basemen and he’s my 46^th overall hitter, which is pretty in-line with his current ADP. The Athletics love to platoon hitters, but Olson is going to play every day. His bat and glove are too good not to be in the lineup. Anyway, based on my projections I don’t think it’s crazy that Olson is going where he’s going. What is crazy is for someone, especially expert(s), to say it’s ludicrous to draft him at his current ADP.

Maybe I’m being an internet troll but at the very least this is a podcast I can no longer listen to. Maybe they’re right and I’m wrong. I can no longer listen because our method of evaluation is completely different.

Like everything I post this is going to be another stream of consciousness. In the last two seasons I’ve won three out of four NFBC leagues. What all those teams have in common is my hitting was exceptional and my pitching was mediocre.

After reading books by Robert B. Cialdini, Richard Thaler and of course Daniel Kahneman I know one of the hardest things for humans to do is challenge our deeply held beliefs especially when we’ve been successful.

I’ve been above average at finding pitching off waivers, but the  starting pitching I draft always struggles. In the past the most wins I would project is 16 with the majority of pitchers getting 12-14 wins. The reason why I didn’t have much of a difference (or spread) in wins is because wins are unpredictable. In the last two seasons Jacob deGrom has 21 wins while Marco Gonzales has 29. Assuming health, what are the odds Gonzales has more wins in 2020? With no offense to Gonzales, deGrom is a vastly superior pitcher so the smart money should be on deGrom.

The data I am going to site are pitchers who threw at least 150 innings per season during these seasons 2015-2019. I choose these years because this is the time frame when starters’ workloads have decreased substantially on a per start basis.

On average 77 pitchers will throw 150 innings. I currently have 134 pitchers projected to throw that many innings. That means only 57% of my projected pitchers will get to the innings threshold (wow!).

Of the 77 pitchers, roughly six will have 18 or more wins (table/image below).

Giving someone a projection of 18 wins gives that pitcher a big advantage against a pitcher who may only be projected for 14. However, if Gerrit Cole makes 32 starts he probably has the highest probability of being of those six pitchers with 18 wins.

I think the six best teams are (in no particular order): Braves, Dodgers, Twins, Yankees, Athletics and Reds (Even before we learned the Astros were cheating dochebags, I thought they took a big hit with the loss of Cole, Will Harris, Wade Miley, Colin McHugh and Robinson Chirinos.)

I don’t want this to happen but suppose Freedie Freeman or Matt Chapman miss most of the 2020 season. Both of their teams heavily rely on them and if they miss time their teams will probably not earn as many wins; in fact it would be a 5-10 game swing. So maybe I project 15 wins for Mike Soroka while only 12 for Chris Paddack. I think Paddack is the better pitcher and will provide better rate stats, but Soroka probably will earn more wins because the Braves should be better than the Padres.

The image below shows the average xFIP and win totals. The data suggests that in general better pitchers earn more wins, but at the individual level wins are still random.

deGrom is one of the four best pitchers in baseball. You could make the argument he’s the best entering 2020. Despite his talent, he still has to pitch for the Mets; behind that defense and with Wilson Ramos behind the plate. I find it difficult to project more than 14 wins. Odds are Cole, Max Scherzer and Justin Verlander have more wins. In the past I would have projected these four pitchers to have between 15-16 wins each because I didn’t want wins to severe alter my rankings. I now realize this was most likely not optimal.

Lets look at two more pitchers: Jose Berrios and Noah Syndergaard. If you remove wins Syndergaard is the better pitcher, but when wins are added Berrios becomes more valuable. My thesis is the Twins could be the best team in the AL, which means if Berrios makes 31 starts it would be difficult for him not to win 14 games. In the past I would predict 13-14 wins for Berrios on the basis that skill is more predictive of wins.

Berrios is currently my 26th starting pitcher. Lowering his win projection from 15 to 14 drops him to 33rd. So, win totals have a pretty dramatic effect once you move past the top 7-8 pitchers (I feel like this is a duh moment). 

Again I go back to the question. Should fantasy draft pitchers with better skills or pitchers with somewhat less skills but has a better opportunity for wins?

The Rays won 96 games and only one pitcher had more than 10 wins. The Braves won 97 games and only one pitcher had more than 14 wins. The Dodgers won 106 and Kershaw had 16 wins, Ryu and Buehler won 14.

Wins are difficult to predict. Pitchers on good teams have a higher probability for wins, but it is not guaranteed. Therefore, I believe the best way to project wins is to stay within a range of 10-16 wins with most in the 12-15 range.