Yes, this is a mistake.
I used F=800 randomly choosed for an example. The Elo-calculating script that i got from an FIDE chessclub uses Point difference 735 to set expectation to 1 (cause the difference with higher Diff is only 3 or 6 didgits behind decimal point)
So at the sources ppl wrote 735 is F. But actually its 400. Elo suggested 800. F can be choosen free - its just a scaling/resolution factor.
One aim of the Elo-experiment here is to figure out if a different F is better for pop. Currently F is 400.
I think a higher F might be better cause its spreads elo min to max range more.
Thanks to MR-TaNk for having found this!
Good to know to have ppl who read stuff and check to find mistakes, errors and misunderstandings.
And we figured out some interesting things when we were checking
"Nehe'z szu:le's volt"
Btw: the original name of Arpad Elo in hungarian language is A'rpa'd Imre E'lo''
And it sounds much different there