I used Dragon Cruncher’s Munch ability (50% chance to devour dragons) on 10 enemy Dragon troops (without Indigestible, Impervious, Fortitude, etc.).
It failed 8x and succeeded 2x.
I used his Munch ability 10 more times against Dragon troops.
It succeeded 7x and failed 3x.

While the success rate in the long run is ~50%, the short term “chance” is not 50%.
It’s 25% half the time, and 75% the other half.

This is not fun for me. I don’t get that edge-of-my-seat gamble of heads-or-tails.
Instead, I get: “Oh, it failed. And again. It failed yet again? Why does it keep failing? 50% chance, my $%! Finally it succeeded! Another success? Nice. Wait, again? It keeps succeeding?? This is broken…”

@Ozball explained to me a while back about how probability works in the game.
Hopefully I can remember it right.
Think of it like flipping a quarter with zero record of any flips before it or after it. So every time you flip the quarter, it’s a 50% chance… Even if 3 times before it, you flipped a heads.
The 50% chance works the same way in GoW.

There’s a Master Chance Program that grants a success to the first user, a fail to the next, and so on. Users queue up before the MCP, and are at the mercy of their position in line. If they had gotten in line a second sooner or later, maybe their outcome would have been different. Are some users getting lucky with their timing and hogging all of the success at the cost of others?

Theoretically you can have a 90% chance and miss it 5 times in a row. If you want a real statistic, then you should note the stats for 100 casts. I am sure it’s around 40-60% then

Comparing a long and short-term is not really realisitc, because the shortterm is part of the longterm. This means, that in a longterm you have probably a spree, where you devour 7 times in a row, but on the other end you have exactly the contrary. So if you got in that negative-spree, then it’s not really realistic to see only this term.

I would say, you had bad luck on your matches and I am sure you will have luck in future ones

No, the game does not communicate every move to the server so this cannot be the case.

It’s definitely true that RNG in this game is streaky, which is extremely frustrating for players. Others understand this better and have discussed it in length.

The way probabilities for independent events work is that they get closer and closer to the given chance when you have more and more trials - in other words, the success rate is the long term rate. If you have a 50/50 chance, then 3 runs in a row could be damn near anything. 10 runs in a row might be five As and five Bs…but probably not. 100 runs in a row would probably be closer to fifty As and fifty Bs, but almost certainly not exact. A billion runs in a row would probably be reeeeeally close to half a billion As and half a billion Bs - and if you “zoomed in” to any given run of three in a row inside that set of a billion, then those three would be right back to the “damn near anything” that we started with.

What you’re asking for is kind of weird. Think about it.

Let’s say the game has a probability fairy and guarantees a 50% rate. The algorithm would be, “If I devoured last time, don’t devour this time.” Now imagine if that would make you feel better.

It means there are turns where you know, with 100% certainty, if you use the skill it will not devour. That means you’re going to have to waste a turn if you ever want it to devour again. You won’t like that. There’s going to be some match where you’d really like a chance to devour twice in a row, but tough cookies you can’t.

Worse, imagine how a 7% chance would work. It’s really exciting if Megavore procs, but imagine if the game is carefully regulating the probability such that if you’ve seen 7 procs, you’re going to have to make 63 matches before the next one. Does that sound fun?

That’s the deal with probability. You’re always hoping to hit the “good” end, where you’re running above 50%. It can happen and it makes for fun stories. It’s not so fun when you hit the equally likely “bad” end, where you run below 50%. If the probability is enforced, then you can never hit either end but overall it turns out things aren’t as fun that way.

A lot of people would rather see no random-based mechanics at all. Most strategy games try to eliminate completely random probabilities. (There are some that depend on it, yes, but the popular ones tend to let you control the odds. For example, in Risk, it’s your choice if you want to take a 50/50 battle or wait until you have enough troops to have an advantage.)