Two Graphs, One Player: The Carson Kelly Conundrum

In the wild and crazy world of Ottoneu fantasy baseball, players can battle it out in a points-style format that is so true to modern hitting that batting averages are meaningless and home runs are more valuable than a 12-pack of eggs. Don’t worry, this article applies to all leagues, regardless of format. Has this ever happened to you?

You’re rostering a player like Carson Kelly, and you see things like high roster rates, high player ratings, and high start rates. You added Kelly when he was go-ing-off! In his first 11 games, he slashed .419/.578/1.097 with a 1.675 OPS and six home runs. “I’m a genius!” you roared from your apartment rooftop in early April. Things were going great.

Until a week ago, Carson Kelly was a 10+ points per game player in the Ottoneu FanGraphs points format. That may mean nothing to you, but you can contextualize it by knowing that a replacement-level catcher is around 3.6 points per game. That’s Tomás Nido. Unfortunately, this early in the season, there are plenty of underperforming catchers like Yanier Diaz (3.5), Adley Rutschman (3.4), and Salvador Perez (3.3) on fantasy rosters. So, a high-performing Carson Kelly in this day and age is like taking your kid to a baseball game and finding no line at the funnel cake stand in between the 6th and 7th innings. But it’s that P/G metric that appears next to a player’s name in the Ottoneu format, and because of that, highly influential on whether a manager puts that player in the lineup each day.

Here’s how points are calculated for hitters:

Points = (AB x -1.0) + (H x 5.6) + (2B x 2.9) + (3B x 5.7) + (HR x 9.4) + (BB x 3.0) + (HBP x 3.0) + (SB x 1.9) + (CS x -2.8)

Notice how valuable home runs and extra-base hits are in this format, and Kelly recorded nine in March/April. That’s the makings of a high points per game type of player. Yet, a knowledgeable baseball fan should be pessimistic, claiming that the real truth is revealed in larger samples. For example, Kelly’s max P/G finish occurred in 2024 when he reached an end-of-season, or full-season, 3.57 P/G mark. While a metric like P/G next to a player’s name on your lineup display can be helpful, it can also be deceiving. Should you start or sit Carson Kelly? Well, he’s an ~8 P/G player! Start him! But a closer look might make you question that move:

Kelly is still riding the high of a career start to the season, but hard times come for us all, and Kelly’s .360/.507/.840 April has turned into a .192/.300/.346 May. Still, next to Kelly’s name is a misleading 7.8 P/G mark and unless you’re an every game viewing Cubs fan, you likely won’t detect that downward trajectory for a few weeks, dare I say, months! Thus, we see the issue with a concentrated metric like points per game, particularly in small samples.

Now, bring in Kelly’s FanGraphs player card:

Let’s start with the red line at the top of the bars in the visual above. That represents Kelly’s season-so-far value. It’s at $17.90. Not bad, Kelly, not bad. But the purple line at the bottom of the bars represents his end-of-season predicted value as generated by Steamer before the season began. Notice how far apart they are. Bouncing between those two lines are the gray bars (plate appearances per week) and the dotted green line (dollar value per week). So, yea, Kelly had a pretty poppin’ weeks two, three, and four. We can forget about week five and bounce right on to week six where the good times kept rolling. But Kelly came back down to Earth and his pre-season expectations in weeks seven and eight. Breaking apart performance into smaller chucks, like weeks, as is done with these player cards, might help you realize the good times are on hold for Kelly and he needs to come out of the lineup.

Just like there are the overperforming Carson Kellys in fantasy baseball, there are the underperforming Adley Rutschmans. Remember from before that Adley is a 3.4 points per game player. See that mark next to his name, and you think, bench. But, take a look at the O’s backstop’s player card:

Rutschman is drastically underperforming his end-of-season expectation, but his playing time (grey bars) remains steady. That doesn’t mean now is the time to move him from the bench and into the lineup, but it does mean the future has to be better. As the old adage goes: buy low, sell high. Where Kelly’s season mark is way above his projected mark and things are beginning to negatively regress, Adley’s can only be expected to go up, positively regressing.

I know what you’re thinking, and yes, a Carson Kelly for Adley Rutschman trade is far-fetched. Still, viewing a player from a fuller picture allows you to make better decisions. Now, let’s all imagine a world where FanGraphs members can save the player cards for all the players on their fantasy roster in one place, and scroll through each whenever they make their sit/start decisions. Imagine dropping a player on your fantasy team, going to your player card dashboard to do the same, and then adding in a new player card for whoever you claimed in their place. Good idea? I’ll put in a good word with upper management for a feature request.

Through it all, Kelly is still a +7 P/G player. But isolate his performance to weeks seven and eight, and he’s below replacement level. You will see that overall P/G mark come down, but it will happen slowly, and winning your league means noticing the downward trend at the top of the hill. You can do that with FanGraphs player cards.