Spell Training

[Ace High]

I don't know if anyone else has noticed or not, but the amount of time it takes to train spells has gotten considerably longer lately. Has anything changed to where spamming spells no longer works, or what, because I just spent the last hour a half running scripts for both dispel magic and faerie fire and didn't gain a single thing that whole time.

And before you ask, I have max int of 21, good luck, master explorer on quests, etc...

[Burzuk]

[Louis]

bada bing bada boom

[Burzuk]

I should also add that the last time we changed anything that affects training rates was way back when we scaled quests down to being slightly stronger than when I was a mortal (instead of being WAY stronger), and yet I've seen people asserting that they're sure it's been changed on many occasions since then (and, conveniently, they're sure it was always for the worse). I don't notice anyone gloat about their good luck when they hit a nice lucky streak with improvements, while only bringing the issue up instead when they train for a few hours and don't get as much improvement as they think they deserve.

In my day, we blamed the random number generator when things didn't go our way. Nowadays it's becoming fashionable instead to make grand sweeping statements about how "things have been different lately" (cue conspiratorial music) and put force broad assumptions based on what is, statistically speaking, a rather small sample size when you consider the percentages and other numerical factors involved in improvement rates.

Here's a trivia question for you math people out there to illustrate what I'm talking about: if you were rolling a 100-sided die, how many times would you have to roll it to be 99% confident of landing on 100 at least once? What about 99.99% confident? The answer may surprise you non-math types.

This will illustrate why the more "sure" you are that the odds "should" fall a certain way, the less accurate your prediction will actually be. Making assumptions about changes in probability without having done a lot of empirical testing is a very dangerous thing. And statistically speaking. you should be far more happy to see multiple skill improvements bunched together over a short period of time than being unhappy to see nothing over a long period of time. When the odds are low, that's the way the numbers work.

[E-ant]

[09:31] <xanthas> every roll has 1% chance of being 100
[09:31] <linkdead> and a 99% chance of not being 100
-
[09:41] <linkdead> 0.99 = 1 - (0.99)^n
[09:41] <xanthas> ummm, who cares, i will still pk burz irl, probabilities are not going to save him
-
[09:51] <linkdead> its 458
[09:51] <linkdead> for 99%
[09:53] <linkdead> time try99.99%
[09:53] <linkdead> 46,049 rolls

[Solitare]

[Mendek]

Actually, each role is independant of the last... so it's always 1%.

[Fishpaste]

Uhh... someone obviously failed stats, or possibly english comprehension.