The difference between DOTA2 RNG and COH2 RNG is Valve controls its RNG. And to make sure RNG doesn't break the game balance put mechanisms to limit it. It is as simple as putting timer or count control on RNG actions.
You know like, side armor have 50% chance to be rear or front armor. So you put a count with timer on it. If RNG's god decides to take front armor value to a 1st shot, automatically the next shot withing 5 seconds will take rear armor value.
And each time it takes front armor value, again and again.
Or like you set an incremental % value, if the 1st shot is front armor value, so the next shot in the next 5 sec has 70% chance to be rear armor value and if not, it goes to 90% etc...
And of course you adapt your overall units values base on those mechanisms. (numbers above are purely random, please don't come and argue on it)
And so you have RNG and fun but you control its upper and lower limits. And you avoid stupid 3 or 4 side shots bounced in a raw because it took front armor value every time and completely break the game.
That's not exactly how it works. PRD is generally implemented so that the first proc chance is actually lower than the desired average value.
http://dota2.gamepedia.com/Pseudo-random_distribution
For example, Slardar's Bash has a 25% probability to Stun the target. On the first attack, however, it will only have an ~8.5% probability to bash; this is its PRD constant C. Each subsequent attack without a bash increases this probability by 8.5%. So on the second attack, the chance is 17%, on the third it is 25.5%, etc. After a bash procs, the probability resets to 8.5% for the next attack. These probabilities average out so that, over a moderate period of time, Bash will proc very nearly 25% of the time.
Furthermore, there are still elements of pure randomness, such as bash, evasion, Ogre Magi's Multicast mechanic, Faceless Void's Backtrack, etc.
i'm a little confused here. if you're talking about the first game then i won't comment because i haven't played it. if you're talking about CoH2 then it's false as games are decided by RNG. it's not a majority, or even a plurality of them, but it certainly happens.
I'm talking about both games. If a player puts himself in a situation where he needs to rely on random chance to win a game, he's fucked up somewhere along the line. Every time you rely on a single roll of the dice, you're taking a huge risk. Good players know that if, over the course of a game, you take only those risks that have a positive chance of resulting in a favourable outcome, you won't have to rely on a single instance of RNG luck to win you the game.
Games aren't decided by single decisions, they're decided by the culmination of all decisions in the game. If you're focusing on a single instance of bad luck and blaming that on a loss, you're neglecting to consider the several hundred other decisions made throughout the game that ultimately brought you to a situation where you were forced to rely on luck to win.
There's no such thing as a 100% chance to win a game, especially when you're talking about a game between two players of similar skill. All you can do is give yourself the best possible chance of success. That's why best-of-1 games in tournaments and single ladder games don't really tell you anything about the relative skill of two players, because in a single game anything can happen. That's just how competition works. It has nothing to do with RNG.
I'll give an example from CoH1, because that's the game I know best. In CoH1, snipers only had 50% accuracy against regular infantry targets that were moving, including other snipers. That meant a countersnipe attempt against a moving target had a 50% chance of success. 50% chance isn't very good when you consider the cost and game impact of a sniper, yet many lower-level players relied on countersniping and got very angry when they missed, and blamed their losses on RNG even though it was plainly clear that countersniping was unreliable.
Good players, however, realized this and adopted different tactics. Rarely in high-level games anymore do you see players building snipers and hiding them while they wait for an opportunity to countersnipe. Instead, they realized that the best thing to do was to use your own sniper as a deterrent, draining manpower at a similar rate to the opponent's sniper and using tools like flanking, vehicles, and artillery to limit the enemy sniper's effective time on the field. If you can keep your sniper at the front for more total time than your opponent's, you're going to kill more infantry and drain more manpower, which is the primary use of a sniper in the first place.
In adopting this style, players were able to avoid the inconsistency that comes with relying on low-probability RNG as a cornerstone of your gameplay. By controlling the sniper rather than going all-in on a countersnipe, players were able to rely on their positional and tactical skills instead of putting the outcome of the game on a random element that was out of their control.
There are always alternatives to consider when dealing with random elements in a game. The best players will find ways to minimize their exposure to high-risk-high-reward plays and naturally gravitate to the "safest", most consistent style possible.