Chance to learn spell % is BS

ankheg

I am sure that it is well known but this number was never accurate. I mean come on, with 85% less than half of them are successful.

Awong124

Yeah, Neera supposedly has 75% chance to learn a spell, but she almost never learns it on the first try.

[Deleted User]

The user and all related content has been deleted.

ankheg

You can also change the difficulty level. It never worked properly in IE engine as far as I know.

Space_hamster

Maybe you guys are just really unlucky with your dice rolls...... drink a potion of genius first.

ankheg

Unlucky? All these 14 years? I doubt it. I can count it if you want. (and I never use potions)

Dee

I don't think I ever failed to write a spell.

Remember that it's not 85 out of every 100; it's an 85% probability with each attempt. It's sort of like the chances of rolling an 18 or higher on your attack roll; it seems improbable, but it's just as likely as rolling between 9 and 11.

(I've also edited the thread's title; a little profanity in the thread itself is one thing, but broadcasting it to the forums is another.)

Awong124

Aosaw said:

Remember that it's not 85 out of every 100; it's an 85% probability with each attempt. It's sort of like the chances of rolling an 18 or higher on your attack roll; it seems improbable, but it's just as likely as rolling between 9 and 11.

If it's 85% probability on each attempt, then he should be able to write it the first time more often than not. He should be able to write it the first time 85 out of 100 attempts. But this is clearly not the case.

Deltaslayer

Aosaw said:

I don't think I ever failed to write a spell.)

This. I don't understand when people say they failed to copy a spell, it always shows in the spell book no matter the spell level.

iKrivetko

Awong124 said:

Aosaw said:

Remember that it's not 85 out of every 100; it's an 85% probability with each attempt. It's sort of like the chances of rolling an 18 or higher on your attack roll; it seems improbable, but it's just as likely as rolling between 9 and 11.

If it's 85% probability on each attempt, then he should be able to write it the first time more often than not. He should be able to write it the first time 85 out of 100 attempts. But this is clearly not the case.

It is possible to have a huge row of failures even with a high chance for success. Besides, I am quite sure that your amount of attempts is far from proving a hypothesis of the chance being lower than stated and you are merely overreacting. Stop being so frustrated about it and better luck next time.

Awong124

iKrivetko said:

Awong124 said:

Aosaw said:

Remember that it's not 85 out of every 100; it's an 85% probability with each attempt. It's sort of like the chances of rolling an 18 or higher on your attack roll; it seems improbable, but it's just as likely as rolling between 9 and 11.

If it's 85% probability on each attempt, then he should be able to write it the first time more often than not. He should be able to write it the first time 85 out of 100 attempts. But this is clearly not the case.

It is possible to have a huge row of failures even with a high chance for success. Besides, I am quite sure that your amount of attempts is far from proving a hypothesis of the chance being lower than stated and you are merely overreacting. Stop being so frustrated about it and better luck next time.

You obviously have either very little or no knowledge of statistics and probability. Sure it's possible, like winning the lottery is possible. If the probability works properly, and the original poster has an 85% chance of success, he has a 0.3375% chance of failing 3 times in a row, which I'm sure has happened to him. And this isn't even a one-time occurrence, it happens over and over again.

Chow

This kind of stuff happens in basically every game with statistics and chances in it - including on the tabletop itself. The odds seem worse than they really are because only the failures stuck in your mind, while the successes disappear.

Awong124

Chow said:

This kind of stuff happens in basically every game with statistics and chances in it - including on the tabletop itself. The odds seem worse than they really are because only the failures stuck in your mind, while the successes disappear.

Even if he fails 3 times in a row once with 85% chance of success is highly unlikely. We are talking about a 0.3375 percent chance of that happening, not simply 0.3375. That's almost 1 in 300. And from what I gather from his post, it's happened multiple times. And it appears this happens to many people. Even if only the failures stick in our minds, having something like this happening 3 times out of lets say 20 times is already too much.

Awong124

Deltaslayer said:

Aosaw said:

I don't think I ever failed to write a spell.)

This. I don't understand when people say they failed to copy a spell, it always shows in the spell book no matter the spell level.

@Deltaslayer

Are you on normal difficulty? The game is set on normal difficulty by default, which means you never fail to copy a spell. I set mine on core difficulty, which has a chance for failure. Higher difficulties beyond that also have chances of failure.

Chow

Awong124 said:

Even if he fails 3 times in a row once with 85% chance of success is highly unlikely. We are talking about a 0.3375 percent chance of that happening, not simply 0.3375. That's almost 1 in 300. And from what I gather from his post, it's happened multiple times. And it appears this happens to many people. Even if only the failures stick in our minds, having something like this happening 3 times out of lets say 20 times is already too much.

How many spells have you tried to write to your spellbook? In fourteen years? For Baldur's Gate 1 and 2, possibly adding Icewind Dales and perhaps even Planescape: Torment, there must have had thousands. Tens of thousands, even.

With that kind of numbers, this sort of stuff does happen too, eventually.

Awong124

Chow said:

Awong124 said:

Even if he fails 3 times in a row once with 85% chance of success is highly unlikely. We are talking about a 0.3375 percent chance of that happening, not simply 0.3375. That's almost 1 in 300. And from what I gather from his post, it's happened multiple times. And it appears this happens to many people. Even if only the failures stick in our minds, having something like this happening 3 times out of lets say 20 times is already too much.

How many spells have you tried to write to your spellbook? In fourteen years? For Baldur's Gate 1 and 2, possibly adding Icewind Dales and perhaps even Planescape: Torment, there must have had thousands. Tens of thousands, even.

With that kind of numbers, this sort of stuff does happen too, eventually.

Honestly, I can't remember from the earlier games. It's not even necessarily relevant, BGEE could have done something that changed how that system works. I seem to recall vaguely that I've never noticed this phenomenon in the earlier games. Lets talk about this game. In my current game, I have probably tried to write less than 50 spells into Neera's spell book. My Neera has a chance of success of 75%. Lets say you are right in assuming only the failures stick in my mind. Even with that in mind, I can remember at least 10 times where it has taken me more than once to copy a spell, and at least 5 times where it has taken me more than 3 attempts. For my Neera to not successfully copy a spell 3 times in a row, the probability should be around 1.5%. For that happen to me on 5 separate occasions I should probably play the lottery more often.

Awong124

I just did an experiment with Neera trying to write Identify into her spell book 100 times. My Neera has a chance of success of 75. I reloaded with each trial.

My results were 56 successes out of 100. I think that is a sufficiently large sample size to statistically conclude that the success rate in the game is BS. Of the 44 failures, there were 5 times where it failed 3x in a row, and once where it failed 4x in a row.

Here is a list of each individual result of my trial, Y=success, N=failure:

YYYNYNYYNN
NYNNYYNYNY
YYYYNNNNYY
YNYNYNYYYN
NNYYYYNNNY
NNNYYYYNNN
YNNYYYYNYY
YNYYYYNYNY
YNNNYNNYYN
NYYYYYNNYN



I thought about the fact that computers are never truly random, and that I might have chanced upon a random seed that favored failure. So I repeated the experiment without reloading. It takes considerably longer because each success meant I had to go into the spell book and erase the spell.

The results were better, 67 successes out of 100, but still well short of the expected 75. I think 100 is a sufficiently large sample size to still conclude there is something wrong with the system in the game. In this trial, there were 4 times where it failed 3x in a row.

YNYYYNYYNY
YNYNYYNYYY
YNNNYNYYYY
YNYYNNYNYY
YYYYNNNYYN
YNYYYYYYYY
YNNNYYNNYY
NYYYYYYYNY
YNYYYYNNNY
YNYYYYNYYN



You are welcome to try this experiment yourselves.

Deltaslayer

Awong124 said:

Deltaslayer said:

Aosaw said:

I don't think I ever failed to write a spell.)

This. I don't understand when people say they failed to copy a spell, it always shows in the spell book no matter the spell level.

@Deltaslayer

Are you on normal difficulty? The game is set on normal difficulty by default, which means you never fail to copy a spell. I set mine on core difficulty, which has a chance for failure. Higher difficulties beyond that also have chances of failure.

Oh, so that's explain it. I play it on normal difficulty.

Bytebrain

Awong124 said:

I just did an experiment with Neera trying to write Identify into her spell book 100 times. My Neera has a chance of success of 75. I reloaded with each trial.

My results were 56 successes out of 100. I think that is a sufficiently large sample size to statistically conclude that the success rate in the game is BS. Of the 44 failures, there were 5 times where it failed 3x in a row, and once where it failed 4x in a row.

Here is a list of each individual result of my trial, Y=success, N=failure:

YYYNYNYYNN
NYNNYYNYNY
YYYYNNNNYY
YNYNYNYYYN
NNYYYYNNNY
NNNYYYYNNN
YNNYYYYNYY
YNYYYYNYNY
YNNNYNNYYN
NYYYYYNNYN



I thought about the fact that computers are never truly random, and that I might have chanced upon a random seed that favored failure. So I repeated the experiment without reloading. It takes considerably longer because each success meant I had to go into the spell book and erase the spell.

The results were better, 67 successes out of 100, but still well short of the expected 75. I think 100 is a sufficiently large sample size to still conclude there is something wrong with the system in the game. In this trial, there were 4 times where it failed 3x in a row.

YNYYYNYYNY
YNYNYYNYYY
YNNNYNYYYY
YNYYNNYNYY
YYYYNNNYYN
YNYYYYYYYY
YNNNYYNNYY
NYYYYYYYNY
YNYYYYNNNY
YNYYYYNYYN



You are welcome to try this experiment yourselves.

Percentages of chance doesn't work the way you seem to think.
You could toss a coin in the air 500 times in a row. Now, your logic says that it ought to be a 50/50 chance of landing on heads, but in reality, you could have 450 tails if you did it.

The percentage of chance is reset for every new try.
As far as your experiment, you had more successes than fails, so it's absolutely in the realm of a 75% chance.

If you did it a thousand times, you'd almost certainly wouldn't nail a precise 75% success of reading the spell.

sarevok57

i've played the original games since the beginning of time, and even if you had 75% + chance to learn spell you quite often learned them first try, but bgee was a completely different story, with 75% which is common for npc mages, i rarely if ever succeed to learn a spell on the first try, so all i do now is pause the game set it to the lowest difficulty write the spell, and then put it back to the hardest, because reloading so many damn times to learn spells is just annoying as hell

Awong124

Bytebrain said:

Awong124 said:

I just did an experiment with Neera trying to write Identify into her spell book 100 times. My Neera has a chance of success of 75. I reloaded with each trial.

My results were 56 successes out of 100. I think that is a sufficiently large sample size to statistically conclude that the success rate in the game is BS. Of the 44 failures, there were 5 times where it failed 3x in a row, and once where it failed 4x in a row.

Here is a list of each individual result of my trial, Y=success, N=failure:

YYYNYNYYNN
NYNNYYNYNY
YYYYNNNNYY
YNYNYNYYYN
NNYYYYNNNY
NNNYYYYNNN
YNNYYYYNYY
YNYYYYNYNY
YNNNYNNYYN
NYYYYYNNYN



I thought about the fact that computers are never truly random, and that I might have chanced upon a random seed that favored failure. So I repeated the experiment without reloading. It takes considerably longer because each success meant I had to go into the spell book and erase the spell.

The results were better, 67 successes out of 100, but still well short of the expected 75. I think 100 is a sufficiently large sample size to still conclude there is something wrong with the system in the game. In this trial, there were 4 times where it failed 3x in a row.

YNYYYNYYNY
YNYNYYNYYY
YNNNYNYYYY
YNYYNNYNYY
YYYYNNNYYN
YNYYYYYYYY
YNNNYYNNYY
NYYYYYYYNY
YNYYYYNNNY
YNYYYYNYYN



You are welcome to try this experiment yourselves.

Percentages of chance doesn't work the way you seem to think.
You could toss a coin in the air 500 times in a row. Now, your logic says that it ought to be a 50/50 chance of landing on heads, but in reality, you could have 450 tails if you did it.

The percentage of chance is reset for every new try.
As far as your experiment, you had more successes than fails, so it's absolutely in the realm of a 75% chance.

If you did it a thousand times, you'd almost certainly wouldn't nail a precise 75% success of reading the spell.

I think it's you who doesn't understand. Certainly it's possible that you can get 450 tails out of 500 tosses, but it's highly improbable. You can try that experiment and see what the actual results are, I'm confident it would be reasonably close to 50%. With a sample size that big I'm confident it would be within 5% at least, if not within 2%.

The probability of tossing 450 tails out of 500 tosses is 7.07x10^-82, that is 1 in 1.414x10^81, that is the order of magnitude of 10 with 81 zeroes after it. You have a much higher chance of winning the powerball, good luck.

I think a sample size of 100 is very reasonable for a two-outcome test. The margin of error certainly wouldn't be greater than 8% from the results of the second test, and certainly not the 19% from the first test.

ankheg

Although I have to admit that I failed my probability exam back in school, I believe @Awong124 is right. And it is quite easy to create a bad random calculation in programming. I will test it later with my own mages (int 18,19) but I am sure I am right.

Dee

There's a difference between mathematical probability and statistical probability.

Statistical probability suggests that the likelihood of a certain outcome diminishes or increases depending on the results of previous trials. In other words, if your first toss comes up heads, then your second toss is statistically more likely to come up tails, because the statistical probability of it coming up heads twice in a row is lower than the statistical probability of it coming up heads and tails in equal turns.

Mathematical probability, on the other hand, means that with each toss of the coin there is a 50/50 chance of either outcome, regardless of how many times the coin is tossed. In other words, you might toss the coin once, and it would come up heads, and the second time you toss the coin there is the same probability of it coming up heads. Which is to say 50%, each time.

What we're dealing with is an expectation of statistical probability, which suggests that for every hundred trials there should be a direct correlation to the percent chance of success. In other words, for an 85% chance of success, there should be 85 successful results in 100 attempts. That's not what's happening.

What's happening is 100 individual attempts, each calculated on their own, using their own set of randomly generated numbers. Which means that each attempt has an 85% chance of success. This likelihood of success doesn't increase or diminish depending on the number of trials. You might have fifteen failures followed by eighty-five successes. You might have no successes at all. You might not have any failures. When it comes to randomly generated numbers, the key word is "random". The game's engine does not create results that take into account all previous results.

Anyone who has played tabletop D&D understands this. You could have a night where you roll nothing but 1's and 5's. That same night, the person to your left might roll a series of natural 20's. It doesn't mean the dice are rigged; it means you had a string of bad luck.

And sometimes you might have several strings of bad luck. Or you might find yourself rolling really well during combat but rolling terribly in social situations (such as writing spells, for instance).

bbear

@Awong and many others: Can you repeat your experiment of scribing scrolls (about 30 more times) in a trial of 100? We can test whether the probability of success is indeed not 75% for Neera or the results of your experiment couldve occurred merely by random chance. For each experiment of 100 trials of scribing scrolls, record the number of successes and failures. Then, we take the average of the number of successes in 30+ trials.

-Assumes the results of the repeated experiment are independent
-Assumes the means of the repeated experiment follow a normal distribution

Hypothesis testing:
Ho: Mu = 75
Ha: Mu < 75

Using normal approximation for binomial distribution:
Mu = Population Mean = np (number of trials*prob of success)
Sigma = Population Standard Deviation = npq (number of trials*prob of success*prob of failure)
N = number of experiments

where q = 1-p, n=100 trials
Mu = 75
Sigma = 18.75
N about 30

Stat test:

Z = (Sample Avg - Mu) / (Sigma / SquareRoot( N) )

(We just need the sample avg.)

Taking our level of significance at 5%, if Z is less than -1.645, then we reject the null hypothesis that the probability of success for Neera's scribing scrolls is equal to 75%; otherwise, we fail to reject the null hypothesis and the results can be explained by random chance. IN OTHER WORDS, if we reject the null hypothesis, we are 95% confident that Neera scribe scrolls less than 75% of the time.

Awong124

@Aosaw

If there is a probability associated with an event, then the mathematical probability will approach the statistical probability (or vice versa) with each increasing trial. So with a sufficiently large sample size, they should be similar. That is more or less the essence of statistics. Obviously if you do one or two trials you're not going to be able to determine if an event has 85% chance of happening.

Statistics is a way of empirically testing mathematical probability. Is it perfect? No, but it's probably the best method of doing it, and with a sufficient sample size it does a damn good job in practice. There is no method that I know of that proves an event's mathematical probability, I don't think it's possible. I've already said this a few times in this thread; failing many times in a row with 85% chance of success is obviously possible, it's just not probable if the system is working properly.

So while you say that it's possible to get a string of bad luck is true, saying statistical probability is irrelevant is false.

Aosaw said:

There's a difference between mathematical probability and statistical probability.

Statistical probability suggests that the likelihood of a certain outcome diminishes or increases depending on the results of previous trials. In other words, if your first toss comes up heads, then your second toss is statistically more likely to come up tails, because the statistical probability of it coming up heads twice in a row is lower than the statistical probability of it coming up heads and tails in equal turns.

That is only a part of statistics, the permutation part. Combination, which deals with separate events that have no effect on previous trials is also a part of statistics. Also, your example is incorrect. Tossing two heads in a row, two tails in a row, head then tail, or tail then head all have the same statistical probability of 25%. Coin tosses are combinations, each toss doesn't affect successive tosses. However, getting a head and a tail once each (with no regard to their order) has a higher statistical probability than two heads or two tails. So if you already tossed a head, it's still equal chance that you will get either a head or a tail on the next toss, you're not statistically more likely to toss a tail.

Aosaw said:

What we're dealing with is an expectation of statistical probability, which suggests that for every hundred trials there should be a direct correlation to the percent chance of success. In other words, for an 85% chance of success, there should be 85 successful results in 100 attempts. That's not what's happening.

That statement is false. Statistics never claims to have "direct correlation" to results. It suggests that with each increasing trial it will get more and more accurate to mathematical probability.

Aosaw said:

What's happening is 100 individual attempts, each calculated on their own, using their own set of randomly generated numbers. Which means that each attempt has an 85% chance of success. This likelihood of success doesn't increase or diminish depending on the number of trials. You might have fifteen failures followed by eighty-five successes. You might have no successes at all. You might not have any failures. When it comes to randomly generated numbers, the key word is "random". The game's engine does not create results that take into account all previous results.

What you're describing is a combination as opposed to a permutation. Like I said before, they are both still part of statistics. If you do the trial enough times, the number of times of success should still be close to 85%. That is the whole point of the number 85%. If what you say is true, then any number from 51% to 99% would be essentially the same. They are all calculated separately, so each trial has a higher chance of success than failure, therefore there would be no difference between 51% and 99%.

Awong124

bbear said:

@Awong and many others: Can you repeat your experiment of scribing scrolls (about 30 more times) in a trial of 100? We can test whether the probability of success is indeed not 75% for Neera or the results of your experiment couldve occurred merely by random chance. For each experiment of 100 trials of scribing scrolls, record the number of successes and failures. Then, we take the average of the number of successes in 30+ trials.

I've done my part already. I'm not doing it again.

Dee

I think that the only way to know for certain if the mechanic is working correctly is to look at the numbers used to determine success. Counting trials is anecdotal evidence, and it doesn't prove anything in terms of whether the mechanic is working as intended or not.

If it's an integer between 1 and 100 compared with a threshold value of 85, then it's working correctly. If it's not comparing with that threshold value, or if the generated integer is between some different number range, then it's not.

Beyond that, the only thing about the mechanic that could be broken is the way the number is generated. And if it's wrong here, then it should be wrong everywhere else in the game (because the same method should be used).

However, if it's not wrong here, and if the mechanic is designed as I described above, then no amount of anecdotal evidence will convince me that what you're experiencing is anything more than a string of bad luck. Highly improbable bad luck, but bad luck nonetheless. Because if the number generator is working correctly, and the mechanic is designed correctly, then the only thing that could cause a deviation is if the numbers just happen to be turning up the same each time.

Awong124

I believe that the mechanics are faulty. And statistical trials don't provide anecdotal evidence. Me or the OP simply saying, "I'm failing more than half the time with 85% chance of success" is anecdotal evidence. I don't expect anybody to be easily convinced by anecdotal evidence either. Statistical trials provide empirical evidence. If you're not convinced by empirical evidence, then there's nothing more to say. That's like saying there is no proof of evolution, and that someone won't be convinced that evolution takes place no matter how much empirical evidence there is that suggests otherwise.

MERLANCE

I scribed 100 magic missiles on a 1st level wizard with 12 intelligence (50%) and got 42 success, 58 failure. I would say thats close enough to say that its about 50%.

Edit: did 100 more with an 18 int mage, 92 successes.

And then INTMOD.2da has these numbers.Column 1 is your intelligence, column 2 is the % to learn the spell

LEARN_SPELL MAX_SPELL_LEVEL MAX_SPELLS_PER_LEVEL MAZE_DURATION_DICE_NUM MAZE_DURATION_DICE_SIZE
0 0 0 0 4 4
1 0 0 0 4 4
2 0 0 0 4 4
3 0 0 0 4 4
4 0 0 0 4 4
5 0 0 0 4 4
6 0 0 0 4 4
7 0 0 0 4 4
8 0 0 0 4 4
9 35 4 6 4 4
10 40 5 7 4 4
11 45 5 7 4 4
12 50 6 7 3 4
13 55 6 9 3 4
14 60 7 9 3 4
15 65 7 11 2 4
16 70 8 11 2 4
17 75 8 14 2 4
18 85 9 18 1 4
19 95 9 99 1 4
20 96 9 99 1 4
21 97 9 99 1 4
22 98 9 99 1 4
23 99 9 99 1 4
24 150 9 99 1 4
25 150 9 99 1 4

Troodon80

I've never really understood why people talk about this like it's broken. I think in all the time I've been playing over the last 14 years, I think I've only had about 50 spells fail all together in about the same number of play-throughs and a few thousand hours.

I'm going to try the Y/N chart up until mid-day (12:00-13:00), since I have nothing much to do today. I've already done two charts and it's almost 9AM. I tend to play on insane, too, so there's no problem with the rules automatically allowing the character to learn spells. I never realised just how long it takes to learn a spell, then erase, and repeat 99 more times. It really is a chore.

So, I'll get to it . For science (and curiosity, mostly curiosity)!

Riolathel

Ya know with 85% probability.. I almost always learn a spell

That kind of pokes a hole in your "Mechanics are faulty" theory.

I also don't think some of you understand statistics.
It is possible to flip 100 heads in a row. Is it probable? No, but the possibility remains..

"omg this coin flipped heads more than half the time. it most be broken."