Only doing recent elections where winner got under 50 percent
No Runoff :
1960 : Kennedy wins as he runs up total in big cities
1968 : Humphrey probably wins as like Kennedy runs up total in cities
1992: Clinton still wins as many clinton voters who only voted for him to vote against bush probably stay
Clinton voters.
1996 : easy clinton victory
2000: Bush probably wins by padding his numbers in Texas
2016 : Clinton wins by one or two points
With runoff :
1960 : Kennedy still wins
1968 : Nixon easily beats Humphrey as Wallace voters in the south probably go overwhelmingly for Nixon
1992 : this is a total toss up as many voters who voted for bush or clinton just so their main opponent doesn't win vote for Perot . So I believe if Perot gets to final round he wins if not clinton still wins
1996: clinton easily wins
2000: this would be very very close and I would say it's a pure toss up but bush slightly wins
2016: Clinton still wins
I agree with all of these except Humphrey winning in an NPV system without a runoff. Nixon was from CA and CA was a very different place back then. Being from CA or TX (or from NY 50+ years ago) is quite a substantial advantage in an NPV plurality system. Nixon would camp out in suburban SoCal running up his margin and IMO he would win by substantially more than his actual PV margin.
Applying this to some earlier elections where neither party had a majority (note that NPV clearly helps the Republicans pretty much every election from the end of the Civil War to the end of WWII due to exclusionary Southern voting laws keeping turnout low):
1948:
Plurality NPV
Easy Truman win. NY or no NY, Dewey was simply too far behind.
NPV with Runoff
This is interesting, because Dewey would presumably flip out and campaign super aggressively after expecting to win outright by 10+ and then being forced into a runoff in 2nd place. Still, the mere fact that Truman + Wallace = >50% should be definite. Truman by 51/49 or so.
1916
Plurality NPV
Still a Wilson win, but probably only by 1.5% or so because Hughes can drive up turnout in the Northeast and Midwest while Wilson's base states in the South were rigged for the lowest possible turnout at the time. Wilson could counter somewhat with NYC and Boston.
NPV with Runoff
Likely a Wilson win. Hughes isn't getting the Prohibition or Socialist endorsements.
1912
Plurality NPV
Easy Wilson win, but slightly closer due to restrictive voting laws in the South
NPV with Runoff
Now
this is interesting. Teddy's base was dedicated and would have a disproportionately high turnout. I have to believe Debs endorses Roosevelt, and plenty of moderate Midwestern Republicans would as well. Of course, Taft would probably endorse Wilson out of spite, but that could actually hurt Wilson with the Northern progressive side of his base. It really comes down to whether Taft's pro-business base stays home or actually votes Wilson. I think most of them stay home, so Lean Teddy Roosevelt.
*This scenario is particularly fascinating to me. If Teddy wins in 1912, and gets his amendment subjecting SCOTUS rulings to overrule by popular referendum (or just gets enough SCOTUS appointees), we would get most of the New Deal 20 years early. He also campaigned on women's suffrage by constitutional amendment. If that were in place in time for the 1914 midterms or the 1916 presidential election, do we still enter WWI? I think Teddy would either get reelected in a landslide in 1916 on his economic and social policy or he would self-immolate over WWI.
1892
Plurality NPV
Obvious Cleveland win, but narrower due to restrictive voting laws in the South. See 1916.
NPV with Runoff
The candidate Weaver endorses wins. I think that would be Harrison if he did endorse. If Weaver makes no endorsement, Cleveland still wins
1888
Plurality NPV
This was close enough that it should still go to Harrison given that he can boost turnout in the North while Cleveland was opposed by the NYC machine and he couldn't boost his turnout in the South, again due to their voting law shenanigans.
NPV with Runoff
Even more clearly a Harrison win IMO because the Prohibition ticket would endorse him.
1884
Plurality NPV
Pure toss up. Cleveland was ahead IRL and the NYC machine backed him this time.
NPV with Runoff
Lean Blaine, because of Prohibition endorsement
1880
Plurality NPV
Lean Garfield
NPV with Runoff
Pure Toss Up
1876 (I know Tilden technically broke 50%, but this was still an EV/PV split and there were major shenanigans in the South without which Tilden almost surely didn't have a majority)
Plurality NPV
Lean Tilden
NPV with Runoff
Lean Hayes, because revelations of what was going on in the South in the 1st round would motivate massive Republican turnout in the Northeast and better enforcement of voting rights in the runoff round.
1860
Plurality NPV
Lincoln win with a split opposition as IRL, but the opposition would likely have consolidated around Douglas early on if winning individual states didn't matter, in which case Lincoln would be a significant underdog.
NPV with Runoff
Obvious Douglas win. Sadly, Lincoln would get crushed 60X%/3X% in a runoff. This election was the best thing the Electoral College has ever done for the country.
1856
Plurality NPV
Obvious Buchanan win
NPV with Runoff
Still a Buchanan win, as Fillmore would endorse, but narrower than IRL
1848
Plurality NPV
Still a Taylor win due to the urban states
NPV with Runoff
An even bigger Taylor win because Free Soil would never endorse Cass
1844
Plurality NPV
Pure toss up
NPV with Runoff
Lean Clay as he probably wins any anti-slavery voters who still turn out.
1824
Plurality NPV
Incredibly obvious Jackson win
NPV with Runoff
Still an incredibly obvious Jackson win. This election was the 2nd best thing the Electoral College has ever done for the country.