It's been proven that the hit rates shown in XCOM are basically bullshit. There's a bunch of hidden modifiers which determine the true hit probability.

I don't remember why it worked, but I remember that it didn't like it when you only took high accuracy shots, and if you missed a lot first, your 90%s usually hit.

The hit percentages in XCOM are a complete lie. I remember one time my sniper up on a rooftop missed a ~95% chance shot on a mook, who then turned around and crit the same sniper with a pistol aimed through two box trucks and cover. I think that's when I quit.

Two things:

1. It wasn't missing on a 90% that was an issue; you could miss shots that showed 100% because it was actually a 99.5% chance and the display number got rounded up.

2. It still doesn't use real probability. They literally programmed it so anything over 50% was actually a much higher percentage behind the scenes to make you feel like you were doing better (and because computers can't do truely random numbers). This is actually a super common thing in all video games that use percentages or dice rolls. And becsuse it's a computer, if you opened up the code to see the starting seed and all the math applied to it, you could accurately predict everything in the sequence.

I feel the pain you clearly just felt

The newer XCOM games actively remember and give you the same result if you reload. To discourage save scumming. I would get stuck in one where I'd miss a 90%, reload, and still miss. Because the next shot was always going to miss. So I'd take a different shot with someone else that wouldn't matter, come back, and suddenly I'd make that 90%.

The original 90's games, you could just save scum. But you'd still miss that 90 sometimes.

xcom percentages are also not how probability works. When you need 95%+ just to have an effective 50/50 outcome, something is royally fucked.

The "modern" XCOM games cheaped out on their game mechanic budget.

The game Phoenix Point made by a much smaller studio with fewer resources came up with a vastly superior way to tackle hit probability. In that game you can free aim using a reticle made of two concentric circles. The outer circle represents where your shot(s) have a 100% chance falling inside of. The inner circle represents where 50% of your shots have a chance of being inside. The more accurate weapons have smaller circles. Then when you shoot the game simulates the path of your shots and any character or environmental object that gets in the way will be taken into account. If you fire a burst or shoot a shotgun, you're not bound to only 100% hitting or 100% missing. You can have a partial number of rounds or pellets hit the target, while others might miss, be blocked, or even hit another enemy or ally if they were sharing that cone of probability.

This makes the whole thing feel far more real than the shitty dice roll system XCOM relies on that just feels cheap and simplistic in comparison especially for a game of that price. Too bad that overall Phoenix Point had difficulty curve issues and the story was not every interesting to me at least.

If they ever make a new XCOM game I really hope they make that mechanic more like Phoenix Point's. And also lose the arbitrary turn limits that they've introduced in XCOM2 because they force a reckless game style that I absolutely hate in those types of games.

I understand that it's just bad luck, but I was still so fucking pissed off every time someone missed while litterally standing infront of someone and dying because of that.

What is XCOM?

I'd assume that the reason I lost all my finding in xcom fairly quickly too was because of bad AI coding, not because I suck

The fairest Gollop accuracy is in the GBA game though.