I’m trying to figure out how much RRF matters and how to rate players defensively and I’m finding it very confusing.

1901 Nap Lajoie has a 5.75 RRF at Shortstop and a 5.92 RF playing that position (supportively across all data since 2008 so not a small n thing)

Meanwhile 2007 Troy Tulowitzki has a 6.50 RRF but a 5.86 RF playing that position (again supposedly across all data since 2008).

I’m guessing what’s going on here is that RRF is an unnormed stat and it norms year by year and Lajoie’s range in 1901 was more exceptional than Tulo’s in 2007?

I’m finding it very confusing trying to figure out how much to value offence vs defence when the numbers for defensive performance are weird and unreliable.

Lajoie’s error rate at SS is 0.928 in 1901 real season but 0.954 in the engine? This stat is normed so I’m even more confused as to why it tracks away from the engine. The only thing I can think of is do deadball era fielders get bonuses when with non-deadball pitching?

Probably Normalized and maybe even takes into account Stadium? If Tulo was in Coors, that could factor in?

Posted by anthatortc on 6/13/2026 6:23:00 PM (view original):
I’m trying to figure out how much RRF matters and how to rate players defensively and I’m finding it very confusing.

1901 Nap Lajoie has a 5.75 RRF at Shortstop and a 5.92 RF playing that position (supportively across all data since 2008 so not a small n thing)

Meanwhile 2007 Troy Tulowitzki has a 6.50 RRF but a 5.86 RF playing that position (again supposedly across all data since 2008).

I’m guessing what’s going on here is that RRF is an unnormed stat and it norms year by year and Lajoie’s range in 1901 was more exceptional than Tulo’s in 2007?

I’m finding it very confusing trying to figure out how much to value offence vs defence when the numbers for defensive performance are weird and unreliable.

Lajoie’s error rate at SS is 0.928 in 1901 real season but 0.954 in the engine? This stat is normed so I’m even more confused as to why it tracks away from the engine. The only thing I can think of is do deadball era fielders get bonuses when with non-deadball pitching?

Fielding is normalized against the era of the pitcher on the mound at the time. Fielders will make more errors with a deadball pitcher from 1908 on the mound than they will with a modern guy from 2025.

Real-life range is used to determine the + and - defensive plays, which can be very valuable. Range as a stat from sim games is rather useless. Don't expect the two to match up in any meaningful way.

Fielding is normalized against both the pitcher on the mound and the batter at the plate. This link will help you see the impact: https://www.whatifsports.com/mlb-l/league_averages.asp

RRF is sort of normalized, it’s a custom WIS formula which is why a players RRF won’t match their RF/G or RF/9 on B-R.com. Think of it more like an RRF#. Which is pre-normalized against a historic average pitcher/batter combo.

Fielding is normalized in the game as described above, which is why you don’t usually see large variances on player RRF vs RL RRF since it’s displaying a pre-normalized value. The batter/pitcher matchups that are faced in actuality will still swing it, but since it’s set at a relative midpoint, it won’t swing as wide. FLD for deadball players will often swing wide (better) if they’re regularly behind modern pitchers facing modern batters, and similarly, modern fielders will swing wide (worse) if they’re behind deadball pitchers facing deadball batters.

Fielders with same grade will perform roughly the same with opposite era pitchers/batters.

Okay so I should look at E per G and +/- per G to get a sense of the stats?

Lajoie has 0.182 net plus plays per G
he also has 0.121 errors so he nets out to 0.061 positive outs per game at 2B.

Hornsby has -0.014 net plus plays per G and
0.155 errors per game so nets out to 0.169 negative outs per game at 2B.

positive outs and negative outs are registered relative to a weird baseline here were 0 is 0 errors and 0 net plus plays.

Essentially Lajoie is worth 0.230 extra outs per game.
Even if I assume Hornsby averages 4.5 AB/G his 0.042 extra OBP is reduced to about 0.026 extra OBP after being normed against the pitcher which converts into 0.117 fewer outs offensively. So in this gross oversimplification, 1901 Lajoie is the better overall 2B by about 0.113 outs per game? (Realistically Hornsby closes the gap some with better slugging)

at SS 1901 Lajoie is -0.115 outs per G and 1948 Lou Boudreau is -0.029 outs per G. Lajoie has 0.020 better OBP which norms to around 0.011 net OBP after factoring out the pitcher so he generates 0.050 fewer outs per game offensively.

So from a net outs perspective my best option is to play Boudreau at SS and Lajoie at 2B (if I ignore the impacts of better XBH and higher AVG at both positions)?

Note: I’m assuming significant suppression from elite pitching since this is for a max cap league. Against average pitching the OBPs would not be suppressed which changes the math dramatically. Against poor pitching they’d be inflated which one more makes the math quite different.