The Hidden but Crucial Playing Time Projection

The other day, a discussion related to the Ottoneu Prestige League (which opens again this week) brought out the fact that Willi Castro was the only player with 3B/SS/2B/OF eligibility who was projected at over 4.5 points per game. That was a fascinating fact. Having broad eligibility is really useful, but it is only useful if the player produces and Willi Castro doesn’t produce – but he was projected to produce! Willi Castro over 4.5 P/G is pretty intriguing!

But something didn’t pass the smell test.

In 2024, Castro put up 663.4 points in 158 games for 4.199 P/G. As I said above, he doesn’t really produce. And now I am hearing a message that Castro is projected to increase his P/G enough to be a solidly useful player! That is great!

Except Castro’s projections aren’t better than his 2024 production. In 2024 he posted that 4.199 P/G off a .317 wOBA. Not one of the projections on his player page is better than a .317 wOBA. And given that wOBA and FanGraphs Points scoring are both based on linear weights, they map pretty well to each other. Not perfectly, but better wOBA’s typically mean better points production. But that isn’t what we are seeing.

Every wOBA is down, but four out of six projection systems agree – he will be better this year!

So what can drive increased Ottoneu points league production without increased wOBA?  It is possible for two players to have the same wOBA and different P/G. For example, Yandy Diaz and Alex Bregman both put up .331 wOBA in 2024, but Bregman was the better FanGraphs Points scorer, putting up 5.223 P/G vs 4.903 for Diaz.

This could be because of things that earn points without impacting wOBA. The equation for wOBA accounts for at bats, unintentional walks, hit-by-pitches, singles, doubles, triples, home runs. Fantasy players also earn points for intentional walks and stolen bases, while losing points for caught stealings. Bregman did earn about six more points via theses stats last year, but that doesn’t explain the difference. The biggest chunk of the gap between these two players is just that FanGraphs Points scoring values power a bit more than wOBA. Bregman’s high ISO (.193 vs. 133) made a big difference.

So maybe this is why the projection systems see Castro as a worse hitter but a better FanGraphs Points producer! Maybe increased power and base stealing are driving things!

Well we are maybe getting somewhere. There is no giant change in ISO, which isn’t a huge surprise – none of the systems think he is a meaningfully better hitter so why would they expect a ton more power? But an extra 5 SB with 3 fewer CS adds about .12 P/G to a player’s tally over the season (roughly 18 points over 140 games). That closes the gap at least part of the way, but not all the way.

But the other thing that makes wOBA not match P/G is the denominator. P/G is a per-game statistic while wOBA is a per-PA statistic. While wOBA maps pretty well to P/G, it maps even better to P/PA. We saw the big gap in P/G for Diaz and Bregman, but if you look at P/PA, they are at 1.194 (Bregman) and 1.145 (Diaz). That doesn’t feel closer, but if we multiply both by 4.3 to scale it closer to P/G you get Bregman at 5.136 and Diaz at 4.923. That is definitely closer than we had before.

A chunk of the difference, therefore, is in PA/G. In the last paragraph I scaled from P/PA to P/G by taking P/PA multiplied by 4.3 PA/G but they didn’t put up 4.3 PA/G. Bregman put up 4.37 PA/G vs. 4.28 for Diaz. Even if they had been the exact same player per PA, Bregman would have been the more valuable fantasy player simply by getting nearly an extra tenth of a PA per game.

And that does seem to be part of what is happening with Castro.

ZiPS is living in a different world than the rest, presumably expecting Castro to be lifted for PH or used as a PR or something else far more often, but for all six projections, the significant gain (or in ZiPS’ case, drop) in P/G can be attributed pretty significantly to PA/G. Five of the six projections see a lower P/PA expectation in 2025, with the lone exception being OOPSY, which also sees the biggest gains in SB and the biggest gains in ISO and sees the best overall bat at a .311 wOBA. But even OOPSY doesn’t see Castro as a meaningfully better player than he was – if you merge his 2024 PA/G with his OOPSY projected production per PA, you get 4.209 P/G. That is better than 2024, but by a negligible amount.

So this is what it comes down to: if you only look at P/G projections, you might well come away with the belief that Castro is due for a much improved season. But to really buy into that, you need to believe that he is suddenly going to get an extra PA every 3-6 games.

And that is really the primary point of the article – P/G projections can be pretty heavily swung by PA/G projections and unless you really dig in, you won’t notice PA/G projections. I don’t think anyone was looking at the G and PA projections that Steamer has on Castro (141, 623), comparing them to his 2024 numbers (158, 635) and immediately thinking they looked surprising, but they are!

That PA/G shift is a bigger change in his season-long projection as having seven of his 2B turn into HR. If Steamer turned his 12 HR and 31 2B into 19 HR and 24 2B, we would all notice and have big questions, right?

That shift in PA, of course, can happen. Maybe Castro’s PA/G were suppressed last year by a large number of PH appereances and now he’s a locked on starter? Not really. He had only four PH appearances and one PR appearance last year. That isn’t enough to move him up to 4.12 PA/G, getting much closer to ATC and THE BATX, but neither of those projections called for a meaningfully improved P/G anyway.

And I am not sure he is a locked on starter, either. Roster Resource lists him at DH, but he DH’ed exactly zero (0) times in 2024. He primarily played MI and while I am sure there will be games available there in 2025, he isn’t the starter at either spot. He will be sharing time with multiple players.

Maybe he will hit higher in the order? That is definitely possible, as again Roster Resource has him leading off. But he lead off more than hitting in any other spot last year and he was banished from the top of the order down the stretch, leading off just once after September 2 and hitting 5th or lower in every other game that stretch. Given his bat isn’t particularly good (and projects to get worse!) he isn’t really an ideal leadoff option.

So while my real focus here is just to tell you take a closer look at PA/G or P/PA instead of just projected P/G, I’ll also say – I am skeptical of these projections for Castro. The strong consensus that his per-PA production won’t improve and will probably decline is telling. The fact that some systems are showing scores with big gains in PA/G is, in my opinion, probably just noise.

Which leads to an important question I’ll try to answer another day – is this just a Willi Castro issue or is this more widespread and what should we do about it?