I mentioned Barnett 2x, because no one regards his injury win a real win, I mean the fight lasted less than a minute and nothing happened.
I watched Cro Cops career prior to Pride and K1 so I know the value of his wins very well. And unfortunately many of his wins are not high quality wins. The rankings you bring up don't have much relevance to today because they guys like Fujita, Herring, Randleman, Igor and Waterman ranked where they are is not the same as HW's being ranked top 10 from 2012 to now. Those guys were ranked top 5 or 10 with average records .500 records.
I left those wins out simply because they are not really high level wins or really relevant. I mean beating Aleks back in 04 in his 4th mma fight isn't really a top win, beating Nelson in 2019 isn't really a top win, no one cares about Aliakbari or Yoshida, and that is why I never mentioned Tai as a good win for JDS. If were gonna talk good wins for JDS you go Cain, Mir, Carwin, Nelson in 2012, Stipe, Cro Cop in 09, Werdum, Hunt, Gonzaga in 2010, Rothwell, Lewis, Ivanov. But names like Tai, Struve, Yvel, I don't really mention....Cro Cop has way too many of those so so wins and not enough top level ones.
And to your point about Cro Cop beating older Colmean, Randleman and Igor and comparing that to Francis beating JDS, Andrei, and Cain. The difference is those guys are all top level HWs that have accomplished a lot more than the others have in a more far more developed era.
Cro Cop's win vs Barnett is completely legit. Cro Cop locked and pressured his arms and dislocated Barnett's shoulder or collar bone. Win written on record, and that's the end of it.
Yoshida, Magomedov, Waterman, King Mo, Aliakbari, Nelson were all good wins against fighters who presented significant risk at those times, and who were considered good. Those type of opponents can cause upsets against top fighters, speciall if you fight 6 times per year, they possess skill. Modern UFC fighters would lose to those type of opponents if they had to fight more often without rest, without much preparation, without full camp, fighting with body and mind burn out, and without adjusting peak form just once or twice per year.
A lot bigger amount of those wins really scores in favour for Cro Cop against fighters with low number of fights who avoided risk and stayed inactive.
But even by your criteria and example of JDS count, Cro Cop has: Fujita 2x, Barnett 3x, Herring, Coleman, Randleman, Alex Emelianenko, Vovchanchyn, Wanderlei, Gonzaga, Sakuraba, so he is not behind likes of Cain, JDS, Cormier, and he is not behind them in amount of ranked wins.
Coleman, Vovchanchyn, Randleman were among the most acomplished and decorated fighters in those years 2003-2005 when the fights took place, and were still relevant, so comparison with Ngannou beating Cain, JDS, Arlovski is correct.
Alex Emelianenko entered top 10 rankings just 3 months after Cro Cop fight by beating former big name Carlos Baretto and top prospect Thompson.
http://web.archive.org/web/20041204071256/http://www.mmaranks.com/hw
And he was allready combat Sambo world champion, that means he was more experienced than Mirko because combat Sambo is pretty equal to MMA, it has all elements - striking, takedowns, submissions, and he was training for long time with Fedor and his team.
Even Sakuraba was big favourite in Cro Cop fight, and UFC HW tournament champion who beat 6 UFC HW & LHW champions: Randleman, Royce, Belfort, Shamrock, Rampage, Mezger. So correct historical context has to be respected. Cro Cop's win vs Sakuraba was considered as an upset and pretty big win by all fans, media, journalists in 2002.
HW division was certanly developed enough after the year 2000, and its results and rankings can be considered legit and can be compaired to more recent times. Nobody of those ranked fighters was ranked with .500 record at that time, you are completely wrong, they all had very positive records at those times. HW division has shown the least amount of evolution, and it is the first division that was formed, and first that reached reasonably developed state.