Hit percentage literally makes zero sense

[EGGZxOVERxExZ]

Okay, i tested this so there is something obviously wrong here. My sniper is fully durable and my gun skill is at 15 and perception is at 9

Im at 80 percent hit chance and literally about 15 percent of the time im actually hitting my target, even when it says i have 80 percent.

I know this cause i saved scummed back and forth to see if it was just poor luck or if something is really off.

I'd say probably 3 times out of 10, my weapon actually landed its shot with an 80 percent change to hit,

So....am i missing something here or what? 80 percent should mean a 20 percent chance to miss, Not an 20 percent change to hitZ

I love this game and this is my first gripe after 200 hours.
Im js, its pretty irritating.

[bankster]

I could be wrong but if your Guns skill is 15 you are at minimum 35 skill points behind the skill requirment for using it wich must somehow effect outcome, right?

So what if the numbers you are seeing are only true for times your character sucsessfully rolls some kind of weapon handling roll or something behind the screen?

Im not exactly sure how all of it works...

Here are some previus discusions on the topic of how weapon skills behave against Dodge/Evasion and stuff: https://underrail.com/forums/index.php?topic=3285.0

For that topic above Epeli made this: http://underrail.info.tm/junk/precision.html

Also the Precision page from the wiki: https://www.underrail.com/wiki/index.php?title=Precision

[EGGZxOVERxExZ]

Originally posted by =AC= bankster:

I could be wrong but if your Guns skill is 15 you are at minimum 35 skill points behind the skill requirment for using it wich must somehow effect outcome, right?

So what if the numbers you are seeing are only true for times your character sucsessfully rolls some kind of weapon handling roll or something behind the screen?

Im not exactly sure how all of it works...

Here are some previus discusions on the topic of how weapon skills behave against Dodge/Evasion and stuff: https://underrail.com/forums/index.php?topic=3285.0

For that topic above Epeli made this: http://underrail.info.tm/junk/precision.html

Also the Precision page from the wiki: https://www.underrail.com/wiki/index.php?title=Precision

150, Not 15 >___<

[happybjorn]

Make a record of what happens (a real one, not an impression), then do math with actual data. Literally, I'd say, and because I saved scummed aren't useful information.

[EolSunder]

many old games usually have such bad programming. Either the percentages are calculated wrong, OR the actual percentages they show you aren't programmed to show correctly and there are hidden factors that aren't calculated into the shown percentage.

Believe me, i've played many old games where such is normal, and many have been tested to show that it is broken. If i remember the wasteland games, some online games, many i've personally tested and many others have used the actual game code to show the problems.

This game seems the same, it isn't something you can fix you just have to know its there and work around it. Just today i'm mad at missing stuff (95% chance to hit.. miss, miss, miss, hit, miss, miss). or the 65% chance to hit which should at least be every other shot, but misses 17 in 20 chances, you get the idea.

Just adapt. Use more traps which are 100% chance, save scum until you actually hit stuff and all that.

[Hagmar]

Also to note is that random number generators aren't truly random they been known to favor extremes. most new games just kinda cheat for you so you're less likely to miss 75-99%.I kinda like this as it stops save scumming and is more accurate to what should happen anyways.I think i remember hearing that baldur's gate 3 does this to some extent, exactly how I'm not too sure

[EolSunder]

yea personally its programmed by humans, so i'm guessing if its not intentional just programmers who don't know math or who make mistakes. I think one of the wasteland games breaking down items trying to get rare ingredients one such ingredient had a 20% chance to happen, but in 200 tries it never did. 1 in 5 chance, and after 200 tries nothing. You know the odds of that happening? and it happened all the time.

Again nothing you can do about it, it was either programmed intentionally, or more likely the programmers, devs, etc don't talk to each other and just put in what ever numbers they want. Human ignorance/mistakes usually is the case.

[Qiox]

The displayed hit chance in this game is 100% correct. A thread like this pops up roughly every 3 months or so, for the past 6 years. In each case, it is someone posting either their "feelings", or they actually counted hits and misses for a very small number of shots.

Small samples are of course, statistically meaningless.

The devs posted in Discord the results of batches of 1000 shots vs displayed hit chance and each batch had a number of hits matching the displayed hit chance.

[Hagmar]

People do be forgetting that 80% is missing roughly 1/4 of every shot.It can feel pretty bad don't think Underrail is built in a way where what ever baldur's gate does to avoid really bad luck would work well anyways

[EolSunder]

well the easiest way is to do your own calculations. Of course using the actual game code is the best so you can batch process percentages, but most don't know how to do that.

Take a situation and try it out. Today i tried out a situation where i was stealthed and had a 64% chance to hit a stationary target starting combat. So i went and recorded my first 100 shots to see.

Total shots hit 41 out of 100. Not 64, not close to 64%. 41%. Of course that is a small test but even small tests can start to show you, especially if you notice all game that you have problems.

As for Devs posting, you never believe the makers of the game, there are tons of instances of devs and programmers saying this and that is correct, when it turns out they aren't and they get very quiet after they are proven wrong by players and people smarter than them. Again, many have taken actual game codes to batch process percentages and shown that ingame programming is wrong and they show what the actual numbers are. I never believe Devs and programmers for a second, they always get proven wrong by much smarter people.

Just do your own tests, it isn't hard. Even a small test will show you problems. Most of the time when you take those small tests and make bigger tests, the results don't change.

Proved that in the wasteland games where percentages were broken even though "oh everything is correct and ok". 20% chance to get an item and after 200 tries 0 items. You can simply google calculations to see the odds of rolling a random number 1-5 .200 times, and never getting a 1.

People aren't stupid, if you play a game and notice bad RNG, it usually isn't just made up especially when MANY people comment on it, and start experimenting themselves. 45% chance and you hit 3 in 20? Yea, i believe devs and programmers lol

[Hagmar]

If you did another 100 i would be more inclined to believe something is wrong but with my knowledge of how statistics are your rolls could be actually 64 percent.You really can't predict how luck will affect your outcomes, which is way some games don't let it completely ♥♥♥♥ you over.Its still possible to have bad luck in them just not horrible

[Hagmar]

Looking for more info for baldur's gate isn't bringing up anything maybe im mistaken and its was another game.Wish i could find more in depth info about it. I remember something about the BGs devs talking about how bad rolls felt in EA and how they wanted to or did change how save scumming the game could feel

[Qiox]

Originally posted by EolSunder:

well the easiest way is to do your own calculations. Of course using the actual game code is the best so you can batch process percentages, but most don't know how to do that.

Take a situation and try it out. Today i tried out a situation where i was stealthed and had a 64% chance to hit a stationary target starting combat. So i went and recorded my first 100 shots to see.

Total shots hit 41 out of 100. Not 64, not close to 64%. 41%. Of course that is a small test but even small tests can start to show you, especially if you notice all game that you have problems.

As for Devs posting, you never believe the makers of the game, there are tons of instances of devs and programmers saying this and that is correct, when it turns out they aren't and they get very quiet after they are proven wrong by players and people smarter than them. Again, many have taken actual game codes to batch process percentages and shown that ingame programming is wrong and they show what the actual numbers are. I never believe Devs and programmers for a second, they always get proven wrong by much smarter people.

Just do your own tests, it isn't hard. Even a small test will show you problems. Most of the time when you take those small tests and make bigger tests, the results don't change.

Proved that in the wasteland games where percentages were broken even though "oh everything is correct and ok". 20% chance to get an item and after 200 tries 0 items. You can simply google calculations to see the odds of rolling a random number 1-5 .200 times, and never getting a 1.

People aren't stupid, if you play a game and notice bad RNG, it usually isn't just made up especially when MANY people comment on it, and start experimenting themselves. 45% chance and you hit 3 in 20? Yea, i believe devs and programmers lol

So don't believe the devs.

But instead believe some random person who just typed numbers on their computer instead of providing actual evidence.

No thanks.

[.//slayer]

Take a situation and try it out. Today i tried out a situation where i was stealthed and had a 64% chance to hit a stationary target starting combat. So i went and recorded my first 100 shots to see.

You are comparing a sample of 100 attempts against multiple sets of 1000 attempts.

If you remember the Law of Large Numbers, the more attempts you make, the closer you will get to the true statistical distribution. Now, if you were to repeat your test, say, ten more times and noticed that the results were not getting any closer to the expected 64% chance, then and only then would you have a valid case. Otherwise, it's nothing more than the anecdotal evidence.

[Adam]

This thread was quite old before the recent post, so we're locking it to prevent confusion.