Is it just me or do the percentages seem to lie?

[.doll]

Almost fifty percent of my shots that are a 75% + chance of hitting somehow have a nasty habit of missing. I've even had the three 99% chances all miss. Seriously what's going on here?

[hghwolf]

...I will murder the first person to tote out "That`s XCOM baby".

It`s a mix of bad luck and perception bias.

[Isha]

Get a pen and paper. Write down the odds of each shot and the result for an entire game. Let us know the results.

[.doll]

The most notorious misses I get is practically every 90+ no matter what I do.

[hghwolf]

Because you notice and remember those misses more than say... a 50% miss, and you`re expecting the 90% shot to hit.

If you take a large enough sample, say ten full games, then the hits will average about where the percentages say they will.

[.doll]

I expect a 90% chance to hit more than a 50%, I of course expect a 90% to hit more that 25% but that's just not happening.

[Bindal]

It's you - it was already tested several times that the odds shown are pretty much accurate.
https://docs.google.com/spreadsheet/ccc?key=0An6FkQu13utWdE5hMnFUWURRY2dDVHdYZzJhTERSV0E#gid=0

[hghwolf]

Originally posted by seenargles:

I expect a 90% chance to hit more than a 50%, I of course expect a 90% to hit more that 25% but that's just not happening.

And they do in fact hit more often than the 25% shots, assuming you take 25% shots in an equal number to 90% shots in the first place. It`s a quirk of the human brain, you can see it in any game about risk management.

Try playing Blood Bowl, same deal.

[.doll]

That proof is horribly flawed.

[hghwolf]

...how ? I`ve seen that table, the results are as expected and well within the margins of error when considering the individual sample sizes.

[.doll]

Look at how many shots were taken on most of them. Most fall below 20 shots, which can't be trusted at all since there's more variables involved in it. 100 shots would have been better, but no 11 shots taken.

[paranoid623]

Here is an article that elaborates a little bit on the RNG and graphs a number of shots and their hit percentages. The graph strongly suggests that the actual hit percentages are in line with where they should be. http://sinepost.wordpress.com/2012/11/11/is-xcom-truly-random/

[Ygolnac]

Originally posted by anthroban:

Get a pen and paper. Write down the odds of each shot and the result for an entire game. Let us know the results.

This has been done by another user and there's a thread in this very forum about it. It used 500+ shoots, put in a spreadsheet and made all the math. The result was: RNG does not cheat, all the perchentage where right.

Then there's the psycological factor: you usually don't take low percentage shot, and shot only when you have high percentage. Doing so, every time you miss, it seems unfair, but if you would start taking pot shots (25% or less), you will notice you would hit more times than expected.

[hghwolf]

Admittedly for some percentages the sample size is pretty low. But consider that there are very few percentages with massive outliers, even with the small sample size. It`s not entirely comprehensive by any means but it`s a good indication that the chances are actually accurate.

At 150 shots at 45% accuracy the difference from the perfect outcome is only 5%, which is an acceptable margin. 146 shots at 65% accuracy, 1% difference. The greater the sample size the closer the difference gets to 0%, even then just a few percentages would be entirely enough as it seems unreasonable to assume the game would fake only a specific set of accuracy ratings.

The highest difference from the projected outcome is a whopping 76%... from a sample of three whole shots. That sure is indicative...

[Kelloggs]

The key answer is exactly what the very first reply said....

"Perception bias"

You will remember misses during high chance ... and you will forget hits during high chances, because those are the expected result.

[Isha]

The power of Science!

.. Well math really. But I guess recognizing cognitive bias is possible because of science. Some people argue math is a science. Meh. Read wikipedia article "selective perception".