I'm going to respond to many different people here, both directly and indirectly. This post will be a long one.
First of all, I can't help but pointing out that it is not a 75% chance, but in fact a 68% chance of getting 1 or more KOs (I hope I did that right, I have a final exam in this stuff on Tuesday). But that is besides the point. I'm not saying this specific example is a good strategy, but I do believe it is a viable strategy. Would someone be using OHKO moves if they didn't want to win? No. So while it may not be a good strategy, it is a competitive one.
OHKO moves have a 30% chance to hit. That means that they have a 70% chance to miss. The chance to miss 4 times in a row is 70% * .7 * .7 * .7 = 24.01%, meaning that it has a 75.99% chance to hit at least once. I floored it to 75% for simplicity. The point is that you will kill things quicker and more reliably using more conventional tactics unless you are very lucky, and this has only became more true over time, as Pokemon have developed more new and inventive ways to beat the fuck out of each other.
And that is why OHKO moves are banned. Not for being too good, but for being too random. For most of the OHKO move users, playing to win isn't even a consideration, because they're around 10, and are using them because they think they're cool. That's good for them, but it doesn't have any competitive merits.
If it is not that good, why was everyone complaining about its ability to sweep everything. And even if what you said is true, what's wrong with relying on luck. Some people may not realize this, but Skill is not the opposite of Luck. Skill takes Luck into account, and makes decisions based on it. So if a luck based strategy is good, then skilled people use it. Hence it becomes over-centralizing.
Theorycraft discussions always assume either the best case scenario or the worst case scenario, whichever better supports the person's argument. While it is nice to see a tactic at its finest, or to prepare for the first, the outcomes are not always this extreme.
Inconsistent has a very wide degree of variance. It can result in a team full of Ubers being swept by a
freakin' Bidoof, or it can result in the user being KOed on round 2 without accomplishing anything at all - and it only took that long because they used Protect on round 1. Player feedback has nothing to do with this process.
What's wrong with relying on luck is that it goes against every single aspect of the competitive mindset. It's not a coincidence that competitive players are anti luck, and not just here either. Someone later in the thread brought up "No items, Fox only, Final Destination!" That meme accurately summarizes the mindset at work within the Super Smash Brothers series. It's in a more extreme form than I, and most here are advocating, but the principal is the same. No items = remove luck, same character = remove matchup advantages/disadvantages, non dynamic stage = focus on fighting each other, not the terrain. And that character is Fox because he is generally regarded to be the best character. But moving on.
Luck is the enemy to any competitive player because they don't need, or want to rely upon it to win, and so it can only cheat them out of victories that they would otherwise be entitled to. Even if the luck works in their favor, it still makes the victory hollow.
If you don't believe me, try writing a warstory in which the outcome of the battle comes down to randomness. Actually don't do that, that will just make Haunter use Outrage.
This is the reason why if you talk to the competitively minded, or optimizers, or whatever the equivalent is within a given field, and analyze why they are doing as they do, you will find that there is one thing they all have in common - optimal tactics seek to minimize luck as much as is possible.
Someone else later in the thread brought up Magic: The Gathering... it's not a coincidence that there are redundant cards and cards that let you draw cards and cards that let you search for cards... even within that Marvel vs Capcom style metagame, there are plenty of tools to maximize your chance of being able to pull off your gamebreaking combo of choice, thereby minimizing luck.
I could keep going, but my point should be clear by now.
Several people in this thread brought up that there are those who will simply hide behind the words "uncompetitive" and "overcentralization" to hide the fact that they don't have an argument to support their claims. That's a fair point, but at the same time some people will abuse any system. That does not make a system good or bad. It's better to have clearly defined terms so that people can quickly convey what they mean without having to launch into a pedantic explanation each and every time and then move on to the rest of what they have to say and just call the people out who use those words as the alpha and omega of their argument than it is to have people constantly talking past each other or being overly bogged down by minutia due to not having clear and objective definitions for the terms.
Someone also mentioned the whole Timmy/Johnny/Spike thing, and Scizor replacements.
First I'd like to say know your target audience. There's a lot of competitively minded Spikes, and some Johnnys, but Timmy just isn't that interested in that, since he's only playing for "fun".
And if you banned Scizor, all that'd happen is people would move to whatever is second to Scizor and performs a similar role, and just use that. You've lowered the bar, but you haven't done anything for or against skill.