Profile of MilkaCow

Yes, tanks in this game are sensitive. Please be nice to them and treat them kindly.

+1 for remove

This happens every game with every player.
This is why noobs can beat skilled players.
This is why people quit the game.

People want; When I do X, Y is the outcome.
Compare it with another RTS, Age of Empires: When I get cavalry I can beat those archers, simple as that.
Ok make it a bit more interesting and add SOME Rng, specially with accuracy of arty, but how in god's name can you balance a game that is random generating outcomes?

[Slightly OT, more focused on your style than content]
Interesting point of view, but you are not speaking for all people. This "People want" turns your post from an observation and opinion to exaggeration and propaganda. You want it, I guess quite some people want it, but certainly not everybody.
[End OT]

Actually as much as I tend to curse RNG when it treats me unkindly, I still like the RNG a lot. I think especially for abilities the randomness should be reduced (Heavy engine dmg due to AT-nade), but for accuracy/scatter/penetration it's fine. I like the fact that the outcome is not always certain, but that there is some randomness. Let me elaborate why:
Games like *craft, Age of Empires and CnC are using a rather simple model and the outcome of a battle can easily be predicted. For such games muscle memory and knowing which unit counters which (as in, which Build Order) is most important. Games tend to play out pretty similar due to the rock-paper-scissors system behind units. Strategies tend to be made prior to games and then just followed.
Company of Heroes on the other hand is less predictable. You can on average say how good a unit performs against another, but you cannot say for certain. For some weapons (small arms) the simple amount of shots fired causes the RNG to be only visible on the scale of a small engagement, but over the whole of the game it disappears. If a unit overperforms one fight chances are it will underperform in another. For tank shots the RNG plays a far bigger part, since they will most likely only fire a couple hundred shots per game. Overall, you cannot predict how a fight will end - You can say in 99 out of 100 cases this side will win, but you cannot say for sure. Overall, this makes the game a lot less linear and increases complexity as well as requirements to player skill. While in games like SC2 mostly muscle memory is important, here you need to adapt to the situation need to watch fights more closely, since you cannot predict the outcome. I like that, since it keeps the game interesting (the same fight doesn't always end the same way) and thrilling to watch. As I said prior, there are some parts about the randomness I dislike - The ones with a low chance, high impact like for example flame crits or heavy engine dmg from atnade/faust.

I guess different people just want different things from the game. Maybe find some compromise by reducing RNG (as was already done with grenades), but I would not completely remove it, since it's an integral part of CoH in my opinion.

The M3 armor was nerfed while one of it's main targets (the pioneers) damage was increased. That is what's making the M3 underperform. The video examples illustrate this clearly.

Now without pointing fingers on anyone, saying those videos are stupid is frankly a stupid thing to say. Their purpose is to isolate the unit in a favorable or unfavorable situation and test it's ability.

To think I will use it exactly the same way in the video is wrong.

That is exactly the problem there. The whole system was changed, but you seem to have not adapted. Pios might've been a primary target for a M3 in the old one, but I wouldn't say they are anymore. Also Pioneers are a combat unit now, not 'just' a builder. You can still fight Pioneers, but only on range.
Pio DPS with 100% accuracy:
08 15.435
15: 7.900
30: 2.726

As you can see there is a huge difference in DPS even at range. Full auto units shoot smaller bursts and slower at range, not only with less accuracy. If they are moving the bursts tend to get even smaller.

Great targets for M3 currently are Grens at all ranges. Close range will get you fausted usually, but depending on the situation that is often not bad. Pios on long range work good as well. PGrens are okay - they will deal a lot of dmg, but also bleed quickly. Snipers are easier to kill than ever.

Besides Pios in close range, LMG42 Grens are a dangerous prey due to their higher penetration. I think the biggest problem with the M3 currently is that people drive close. Keep it more on range and only swoop in to finish off, to make him faust you or to push him out of cover if you must. I think it might need some minor buffs and I would like it rear armor to be the same as the frontal one (since it's non turreted and the autorotation is sometimes annoying as hell), but besides that I do not see much problems with the M3 itself, but more with the 222 which is too strong atm in my opinion.

You can get 99% to 100% of 4 Grens in cover, that would be 3,96 to 4 entities. For Cons you can only get 25% to 45% of 6, so 1.5 to 2.7 entities.

I assumed you should at least be able to get 4 Cons into cover if you can get 4 Grens into cover. I assume that disparity arises from the fact that Soviets carry so many Molotov and vodka bottles with them, they need more space.