My russians vs germans isn't "One side pounds" it's "Either side could win, it really depends."
That's the issue. . .RM danger levels do begin to level out to where the average value of OB/DB exceeds the likely modifiers around 120. . .but then you toss in parry and it goes back out the window again. (and surprise, which does that again.)
Based on the motivation of a squad of five 5th level fighters to fanatically kill, fight but survive, or block without hurting. . . .i.e. Full OB / 50:50 / Full DB the results are totally different.
You might be able to assess, that a group of five 5th level fighters intending to block a doorway without hurting the attackers and going full DB. . . .or the danger level of five 5th level fighters intending to fanatically kill the party at full OB. . . .but you can't assess the average danger of "Five 5th level fighters" casually.
Then if you get into specifics of "OK, they only have melee weapons, can we fire missile weapons or spells at them where they're trying to hold the doorway without closing in, forcing them to die in place or leave the door to come after us?" situational and specific modifiers totally change the scope of what's afoot, and the relative danger. . .
I have no problem offering advice to Noob GMs on how to assess danger, but stating "Five fifth level fighters = CR5" is essentially offering an answer so stripped of likely (not possible, likely) variables as to be a deceptive statement. .
Or not. Possibility isn't probability. Sure, a +35 OB/0 DB, AT2, 3rd level orc may kill a +240 OB/+150 DB, AT20, 20th level fighter, but even considering the 20th fighter doesn't have the equipment going with his level, and going with an average parrying (say 140), the orc would need to reroll at least thrice to be able to score a CC. So, sure, you can say that in RM2, he still stands a chance while in D&D he doesn't, but to say that "once you do that, any meaningful CR goes right out the window" would be wrong IMO.
In D&D, it's flatly impossible for the orc to 1 shot kill the 20th level fighter, and the odds of winning a stand up fight are so tiny as to almost equal zero. . .but in RM, while the odds of winning a stand up fight still remain clsoe to zero at this very one sided example of combat, if the orc gets a flat footed, rear, surprise attack on the 20th level fighter. . . .the Fighter has zero quickness DB, zero parry DB, zero shield DB, and eats a 35 + 55 = +90 shortbow attack, which gives a 95% chance of an A critical vs AT1, and a 65% chance of an A crit against AT20, and an A krush 4% of a down/dead reasult and around 25% of multi round stun no parry. . . .thus, in the modifiers and mechanics, there is a VAST difference between the two game systems.
Heck, IRL, commanders would give/assess missions difficulty levels and put against it the necessary men and material (if available). Sure, anyone using a gun could indeed shoot dead the best SEAL if circumstances are right, but most officers would expect their 5 SEAL marines trying to take back a boat held by 2 fishermen to win without any difficulty. A CR is just representative of that: the fishermen are CR1, the 5 SEAL marines are level 10, so their should be able to take down a CR1 encounter without much trouble and spending less than 10% of their resources.
Yep, if the seals get to shoot first, from ambush, they should have a chance of success approaching unity. . .of course, that same Seal commander would say "If my team should get screwed by murphy, and two fishermen with bolt action rifles see the team and shoot first, I might loose some of my men." . . .that's the thing, if the worm turns, and the luck plays out wrong, the odds aren't shifted a little, they can completely invert.
Small team odds are based on minute planning, knowing as many of the particulars as possible, and large scale planning is based on % of acceptable casualties. . .which may disatisfy the PCs if your CR level system assumes 20% of them die every session.