The Percentages are Lies

[Chromius]

I love this game, but I hate when they lie on the percentages to hit. Karmic dice off.

regularly get 4-5 misses with 75/75/85/85/75 to hit

[Caelinus]

Originally posted by ohias:

Originally posted by zero:

except you do get it, cause it occurs naturally every time you roll.

i may not get the number i am looking for, but the odds aren't 0, which is the entire point people keep telling you.

your so worried about time but probability does not care.

I think you misunderstand what probability actually means. Probability is a value to predict a chance of a certain event to occur. Assigning a probability post factum makes no sense. Again, imagine you're rolling a die with an infinite number of sides. The probability of each roll is 0 yet you throw it and some number appears. Just because some even occurred it doesn't mean you could've predicted it and assigned probability to it. The probability is 0 because most numbers form that set will never appear in you rolls.

There isn't enough rolls in the universe to get every roll sequence for 10000 throws, in fact a very tiny teeny number of them will ever appear. The rest are pretty much physically impossible to ever be seen.

None of them are physically impossible to be seen. Until the universe actually dies a heat death all of them are possible, albeit extremely unlikely.

And even then they may show up in other potential universes. We have no idea if reality is infinite or not, so it is possible that all of them *have* been seen infinite numbers of times.

Also no one is assigning probabilities after the fact, we are saying what the probability was of getting the result we got before the fact.

As we approach the end of a particular sequence the probability of that sequence continues to grow until it hits 100%, but before we started to roll it would have been 0.01^10,000.

We are extremely unlikely to predict a chain of d100 rolls even 3 or 4 long, but that does not mean that our prediction before the fact was any less likely than what was actually gotten. We just guessed the wrong one from the bag of <unknowable number> sequences.

[whatdoesthisbuttondo?]

Originally posted by ohias:

There isn't enough rolls in the universe to get every roll sequence for 10000 throws, in fact a very tiny teeny number of them will ever appear. The rest are pretty much physically impossible to ever be seen.

Some might even appear twice or more times.

The point is you can't exclude any of the candidates before you have exhausted all rolls in the universe, and you can't predict when a specific one will appear if it is part of those that do.

This is the fundamental misunderstanding people are falling victim to here, I can only assume this is because our brains desperately want to assign some form of order here where there is none.

[Nomad]

99% chance to hit and miss is not possible. You are rolling a D20 on every attack roll. Nat 20 and nat 1 are automatic hits and misses regardless of the targets AC so the highest hit and miss chance that should be possible is 95%.

Getting chains of luck that poor is possible. Statistically unlikely but possible. Thats why Karmeric dice is on by default. Its supposed to help prevent getting chains of really bad luck or really good luck so the player feels less friction with the randomness of the D20 system.

The D20 system is something I am not totally convinced translates well into video games. As I have said in the past when you are sitting down with your friends playing D&D, the DM isn't really out to kill you they want to progress the story. They wont hold your hand but if you are getting terrible rolls all day/night they may cut you some slack. A video game isnt gonna do that. Rolling poorly can also create some really funny and memorable moments for players playing D&D but in a video game it often just means players get frustrated.

[whatdoesthisbuttondo?]

Originally posted by Nomad:

99% chance to hit and miss is not possible. You are rolling a D20 on every attack roll. Nat 20 and nat 1 are automatic hits and misses regardless of the targets AC so the highest hit and miss chance that should be possible is 95%.

Not that familiar with D&D rules, but is that maybe just accounting for a saving throw of the target that might be "critical success"?

I mean, if doing this properly the game should show your chance to hit, and only after you commit it should roll for the hit, and then if hit roll saving throws, at least that is what I would implement intuitively.

[Timmy]

Originally posted by FatAdvice:

That's just how percentages work.
You could miss a 99% chance to hit 10,000 times in a row. That doesn't mean that it won't average out to 99% in the long run.

WRONG

[Caelinus]

Originally posted by whatdoesthisbuttondo?:

Originally posted by Nomad:

99% chance to hit and miss is not possible. You are rolling a D20 on every attack roll. Nat 20 and nat 1 are automatic hits and misses regardless of the targets AC so the highest hit and miss chance that should be possible is 95%.

Not that familiar with D&D rules, but is that maybe just accounting for a saving throw of the target that might be "critical success"?

Nat 1's auto fail, nat 20's auto succeed.

There is only ever one roll made on some sort of effect, either an attack or a save.

Attacks will always hit on a 20 no matter how high or low you attack is, and always miss on a 1. Saves always succeed on 20, and always fail on 1. So the maximal chance you can ever have is a 95% chance to hit, and the minimum chance you can have is a 5% chance to hit.

[ohias]

Originally posted by Caelinus:

None of them are physically impossible to be seen. Until the universe actually dies a heat death all of them are possible, albeit extremely unlikely.

The number of rolls you need for all of them to be seen is much, much, much, much greater than the time you have until the heat death of the universe. Even if you make one roll per plack unit of time.

Originally posted by Caelinus:

And even then they may show up in other potential universes. We have no idea if reality is infinite or not, so it is possible that all of them *have* been seen infinite numbers of times.

"we don't know therefore I'm right".
How about we stick to what we do know? Just a thought.

[whatdoesthisbuttondo?]

Originally posted by Caelinus:

Originally posted by whatdoesthisbuttondo?:

Not that familiar with D&D rules, but is that maybe just accounting for a saving throw of the target that might be "critical success"?

Nat 1's auto fail, nat 20's auto succeed.

There is only ever one roll made on some sort of affect, either an attack or a save.

Attacks will always hit on a 20 no matter how high or low you attack is, and always miss on a 1. Saves always succeed on 20, and always fail on 1. So the maximal chance you can ever have is a 95% chance to hit, and the minimum chance you can have is a 5% chance to hit.

Yes, I mean the percentage the game is showing, I think I recall seeing 89% a lot on "advantage" stealth attacks a lot.

Again, not familiar with the rules at all, but I'm assuming it is trying to convey more than just the chance to hit in that number, like chance to hit and the target not succeeding in any evasion skill?

[Caelinus]

Originally posted by ohias:

Originally posted by Caelinus:

None of them are physically impossible to be seen. Until the universe actually dies a heat death all of them are possible, albeit extremely unlikely.

The number of rolls you need for all of them to be seen is much, much, much, much greater than the time you have until the heat death of the universe. Even if you make one roll per plack unit of time.

Originally posted by Caelinus:

And even then they may show up in other potential universes. We have no idea if reality is infinite or not, so it is possible that all of them *have* been seen infinite numbers of times.

"we don't know therefore I'm right".
How about we stick to what we do know? Just a thought.

1: You do see them constantly. Every 10,000 d100 rolls is a sequence that had that exact chance before it happened. The fact that the sequences we get seem unimportant to you is a personal perception thing. A sequence that seems more random is not inherently more likely than a sequence that seems ordered, both are equally random. The order is an imposition we put on it.

d100 is a pretty normal roll as it is the basis of our percentile system, there are going to be a LOT of 10,000 long sequences happening just on Earth, all of which were equally unlikely before they happened. So, yeah, it will not take infinite time to see one in the universe, there are probably an uncountable number happening every second in the whole universe.

2: That was a thought experiment, I did it to demonstrate that an arbitrary time limit like the end of our universe is pointless for the discussion. We do not know how many rolls will happen before that, and technically we do not even know if that will happen. We have no idea which sequences will be seen before any arbitrary time in the future is met. We do know that. given infinite time, and infinite number of each sequence will be seen.

[Caelinus]

Originally posted by whatdoesthisbuttondo?:

Originally posted by Caelinus:

Nat 1's auto fail, nat 20's auto succeed.

There is only ever one roll made on some sort of affect, either an attack or a save.

Attacks will always hit on a 20 no matter how high or low you attack is, and always miss on a 1. Saves always succeed on 20, and always fail on 1. So the maximal chance you can ever have is a 95% chance to hit, and the minimum chance you can have is a 5% chance to hit.

Yes, I mean the percentage the game is showing, I think I recall seeing 89% a lot on "advantage" stealth attacks a lot.

Again, not familiar with the rules at all, but I'm assuming it is trying to convey more than just the chance to hit in that number, like chance to hit and the target not succeeding in any evasion skill?

Oh yeah, sorry I misunderstood you. That is exactly what is happening, aside from the evasion skill thing.

If you have a 90% chance of success, that means you have a 10% chance of missing.

So advantage fails only when both land in that 10%, so .1*.1 = .01, or a 1% chance of miss. So a 99% chance to hit. (You roll two d20's for advantage, take the higher, and for disadvantage you do the same but take the lower.)

However, attack rolls do not have saving throws and saving throws do not have attack rolls for a single effect. A spell can have both an attack roll and a saving throw on 2 *different* effects, (as in the spell does 2 things) in older editions, but I am not sure if that is present in 5e.

Basically you should only ever roll a d20 once per effects unless you have advantage or disadvantage, and in those cases you just roll 2, but only keep one.g

Edit for Clarity:

Attacks cause the attacker to roll a d20 against their targets AC (Armor Class). If they meet it or beat it, the attack succeeds and it's affect happens.

Saves cause the *defender* to roll against the attackers Save DC (Difficulty Class). If they meet or beat the Save DC, they effect either fails or has a lesser effect, depending on the spell or effect.

You only ever do one of these.

[whatdoesthisbuttondo?]

Originally posted by ohias:

The number of rolls you need for all of them to be seen is much, much, much, much greater than the time you have until the heat death of the universe. Even if you make one roll per plack unit of time.

The thing is we don't need to see all of them, in fact this whole discussion is about seeing just one.

This is a very real issue in cryptography, where certain algorithms are vulnerable to specific random numbers you might draw from a RN source. While they are extremely unlikely to get, discarding these is not uncommon.

[felmari]

In theory with probability you could still hit any one number every time. The probability does not change the median does.

We have a lot of buffs as well in game that help get within success sequence not to mention karmic dice if one chooses.

[Nomad]

Originally posted by whatdoesthisbuttondo?:

Originally posted by Nomad:

99% chance to hit and miss is not possible. You are rolling a D20 on every attack roll. Nat 20 and nat 1 are automatic hits and misses regardless of the targets AC so the highest hit and miss chance that should be possible is 95%.

Not that familiar with D&D rules, but is that maybe just accounting for a saving throw of the target that might be "critical success"?

I mean, if doing this properly the game should show your chance to hit, and only after you commit it should roll for the hit, and then if hit roll saving throws, at least that is what I would implement intuitively.

On an attack roll a nat 20 is a critical hit. Some DMs will make you roll again to confirm in which case if you roll a 20 again the enemy instantly dies (where it makes sense anyway). This has become less common in 5e but is still around. Regardless a nat 20 on an attack roll whether you confirm the crit or not is an automatic hit. Rolling a nat 1 is a critical miss and sometimes results in other consequences as well depending on the campaign. Maybe you hurt yourself, maybe you drop your weapon ect. Extra stuff like that are typically house rules.

[DadTheNoob]

So, you've not played XCOM2 much, have you @OP?

[zero]

Originally posted by Nomad:

99% chance to hit and miss is not possible. You are rolling a D20 on every attack roll. Nat 20 and nat 1 are automatic hits and misses regardless of the targets AC so the highest hit and miss chance that should be possible is 95%.

Getting chains of luck that poor is possible. Statistically unlikely but possible. Thats why Karmeric dice is on by default. Its supposed to help prevent getting chains of really bad luck or really good luck so the player feels less friction with the randomness of the D20 system.

The D20 system is something I am not totally convinced translates well into video games. As I have said in the past when you are sitting down with your friends playing D&D, the DM isn't really out to kill you they want to progress the story. They wont hold your hand but if you are getting terrible rolls all day/night they may cut you some slack. A video game isnt gonna do that. Rolling poorly can also create some really funny and memorable moments for players playing D&D but in a video game it often just means players get frustrated.

95% per individual roll, but advantage rolls twice, so the odds of getting a number that can hit in that invidiual attack is higher then 95%, even if 95% per roll is still true.

advantage will take the highest possible number and use that to determine.