Ah, the eternal question: "Who's on first?"
Along with designated hitter and the corner outfield spots, first base has historically been on the bat before all else positions. So, not surprisingly, three of the top ten hitters (by era uncorrected consolidated battings stats like OPS+ and wRC+) of all time are first basemen.
While cases can be made for other players, particularly multipositional players like Musial, Mantle and Foxx, the argument for the best first baseman in MLB history boils down to two players: Lou Gehrig, Albert Pujols. It's a tight enough race that I'm going flip my usual format and have the player discussion mostly before making my selection.
First, some raw batting numbers:
Albert Pujols (at time of writing): .325/.416/.609/455 (avg/obp/slg/HR)
Lou Gehrig: .340/.447/.632/493
It's clear that the Pujols's bid for a roster slot depends heavily on correcting for the different eras he and Gehrig played in, as well as their relative contributions on defense. League and era context matters and that adds an addition wrinkle:
Sadaharu Oh: .301/.445/.634/868
Yes, that's 868 home runs. If context didn't matter, the choice for first base would be easy: Sadaharu Oh, hands down.
Since context does matter, we'll have to look at the context.
Some of that is fairly straightforward: wRC+ allows us to compare batters to their contemporaries, but is only available for the MLB players.
Compared to their peers, Gehrig has some advantage over Pujols with the bat. It's not a large difference, though, and league average normalization still doesn't account for the relative qualities of average talent in different eras. Throw in Pujols's apparent advantage in defense, and the fact he plays against talent drawn from a much larger population and I'll call it a wash.
However, wRC+ weights events according to their value in a league average context. For the greatest possible team, the offensive environment is considerably more potent, so not getting out (and getting on base) is an even more valuable skill than normally.
From 1923 to 1938, Gehrig's career before his disease disabled him, MLB average OBP was 0.3448, so Gehrig's .447 was about 1.30 times better than his peers.
Pujols plays in the modern context and across his career, OBP has averaged 0.3304. We put Gehrig in the Pujols modern context, at least compared to their respective peers, by multiplying 1.30 by .3304 to get Gehrig's equivalent modern OBP as .428. So, neglecting the talent difference of the eras (and it's hard to quantify), Gehrig's equivalent OBP is notably higher than Pujols. With the rest of the line up filled by the other greatest hitters of all time, that's suggests Gehrig would be a better choice than Pujols.
But oh, what of Oh? To even begin our comparison, we need a projection of how Oh would have done in the MLB. Fortunately, Jim Albright has already done such a projection, based on the relative performances of MLB players who moved to Japan and explained in part 2 of his discussion of Oh's Cooperstown qualifications. Albright figures Oh's hypothetical MLB career would've resulted in a .279/.412/.484 with 527 home runs line from 1962 to 1980. Clearly, Oh's slugging takes a fairly severe hit, though the 1960s and 70s were low power eras, so the effect is exaggerated some.*
*I'd like to go back and calculate Oh's normalized slugging in an MLB context and make an attempt at estimating his Runs Created value as well. I'd probably use a normalized gross production average (1.8*OBP+SLG)/4 as a first cut, as it should be on approximately the same scale as OPS+ and wRC+. But I've put enough time in this already, for now.
What's more important in this context is Oh's projected OBP compared to the MLB average for those years. Dropping his weaker age 39 and 40 seasons (largely for a better comparison with Pujols and Gehrig), Oh projects to have been about 1.32 better at getting on base than the average major leaguer oh the time. Set in the modern context, that would give Oh an OBP of 0.435, higher than either Gehrig or Pujols.*
*As discussed later, Albright was extra conservative about walk rate. Even if Oh's actual walk rate was used (rather than correcting for the differences in leagues, which would favor Oh), his on base ability projects to about 1.36 times MLB league average, or a .450 career OBP if set in the modern .330 OBP league average context and about .470 in Gehrig's time. It is not unreasonable to conclude that Sadaharu Oh's eye for the strike zone and on base skills were on the level of Wiliams and Ruth*/
And Oh's projections are for years well into the post integration era, so the average talent level they are based on is higher than Gehrig's time, though likely somewhat lower than now. Still, given the offensive context of this potent team, Oh's on base ability gives him the advantage.
By nearly accounts, Oh's defense was well above average. Giving him credit for a modest 5 runs saved above average per season seems reasonable. Added to Albright's projection, that leave Oh with about 80 runs saved above average on defense, making him comparable to Pujols. Even as a league average first baseman, Oh's bat still gives him an edge over Pujols in this line up.
There's also one more card in Oh's hand. While Pujols appears to be leaving his peak performance period, his career line is currently little affected by it. Gehrig died rather tragically had his career and life cut short by his disease and so didn't really have the typical post peak tail. Oh, instead, played for 21 seasons in NPB and demonstrated the ability to play at an elite level into his late 30s.
Selection Sadaharu Oh
The greatest slugger in Nippon Professional Baseball history, with 868 home runs, Oh makes the roster not on his power, but on the strength of his on base ability. His career NPB line of .301/.445/.634 with 868 home runs is not far from what might have expected from Lou Gehrig had he played another 7 years. Jim Albright projects Oh's hypothetical career in the MLB as 0.279/0.412/0.484 with 527 home runs from 1962 to 1980.* Normalizing Albright's numbers by league average, and Oh's ability to get on base (in comparison with his peers) exceeds even Gehrig. In the greatest possible team run scoring context, that's enough. Oh's strong defensive reputation is an additional perk, though.
*Albright was actually conservative in his walks estimates. While he found that players are more likely to walk in the MLB context, he didn't correct Oh's walk totals for either the league difference, actually in Oh's favor, or the extra playing time. Since more walks means fewer at bats, Oh's entire line improves. Just extrapolating Oh's walk rate (at 20.5% behind only Ted Williams's 20.6% among MLB batters) without accounting for the league difference gives a career stat line of 0.287/0.433/0.497
Unlike the Negro league greats who didn't later play in the MLB, Oh (and other Japanese players) actually did play against MLB baseball teams in exhibition games, usually against higher quality teams. His 426 PAs in these games provide nearly a full season's worth of data and help validate Albright's projections: .260/.413/.524 with 25 home runs and 14 doubles from 1960 to 1979. Note, though, that these games were usually held in the smaller Japanese ballparks of the time.
The left handed Oh's best season probably was 1974. His NPB line of .332/.527/.761 with 49 home runs in 130 games projects to .301/.474/.539* with 32 home runs in 162 games in the MLB. That would be good enough to lead the league in OBP (over Rod Carew) by 31 points and place third in slugging, after Dick Allen and Mike Schmidt. As a rough estimate, that's about 192 wRC+. Compare that to Joe Morgan's MLB leading 171. A fantastic peak season.
*Even with Albright's conservative walks estimate
If you like awards and championships: 9 NPB MVPs, 9 Gold Gloves and 11 Japan Series Championships with the Yomiuri Giants.
When I woke up this morning, Gehrig was my selection.
The most legendary of all first basemen, Gehrig, a left handed batter, shined brighter among his peers that even Oh and Pujols. In addition to playing in 2130 consecutive games, the Iron Horse had four seasons with an fWAR greater than 10, a hall of fame worthy achievement in itself. The best of those was probably the Murderous Row campaign of 1927, when Gehrig hit .373/.474/.765/210 (avg/obp/slg/wRC+). Gehrig's optimism and character in the face of a debilitating, terminal disease have justly added to the legend of a truly elite ballplayer.
Gehrig's career line is the best in MLB history for a first baseman: .340/.447/.632/147 with roughly average defense for 125.9 fWAR in only 14 mostly full seasons of play time.
Despite being only league average with the bat this season, The Machine is in the midst of a clear Hall of Fame career. Never accruing less than 6 fWAR in a season from 2001 to 2010, Pujol's current .325/.416/.609/164 line with 88.5 fWAR puts him behind Gehrig in value at first base. While Pujols's peak offensive seasons haven't been quite as valuable as Gehrig's, with only the 2003 season (.359/.439/.667/185) exceeding 10 fWAR, his strong defense (almost 70 runs saved above average) and greater talent pool Pujols has faced make a strong argument for Pujols as a co-equal with Gehrig as greatest MLB first baseman of all time. And he stands well with Rogers Hornsby as the greatest right handed batter as well.