Repairing unique items

[Penchant]

No, you're correct. There are, however, a number of actions in the game for which many players will swear on a stack of Bibles that you can always get much better results if you do them X times for Y hours, instead of W times for Z hours.

In this case, though, there is actually a grain of truth to it. Gathering is less effective when you have more fatigue, and each hour of gathering increases your fatigue. Thus, if you gather for 2 hours, then rest for 2 hours, you may have better results on your next 2-hour gathering session, simply because you will no longer be at risk of getting the "you were too fatigued to find anything" message.

Would you agree doing three four hour sessions of diplomatic actions spent on the same thing is more effective than one session of 12 hours?

[Anaris]

Would you agree doing three four hour sessions of diplomatic actions spent on the same thing is more effective than one session of 12 hours?

I honestly can't see how. I wrote that code myself, and the time is just multiplied into the effectiveness directly. It should scale absolutely linearly (though with a certain amount of randomness, of course).

[Penchant]

I honestly can't see how. I wrote that code myself, and the time is just multiplied into the effectiveness directly. It should scale absolutely linearly (though with a certain amount of randomness, of course).

The randomness is what makes me think that the three sessions of four hours is better because it seems like you are more likely to have the randomness be a benefit as I assume the randomness is also affected by skill.So basically, the several sessions isn't always the best but it is on average IMO.

[Anaris]

The randomness is what makes me think that the three sessions of four hours is better because it seems like you are more likely to have the randomness be a benefit as I assume the randomness is also affected by skill.So basically, the several sessions isn't always the best but it is on average IMO.

Huh? Randomness affected by skill?

The randomness is random. You could end up with one really good session on the 12-hour one, or a really bad one. Similarly, you could end up with three good 4-hour sessions, three bad ones, or some of each. Or all of the above could end up neutral.

It will average out in the long run. You don't get the better of randomness by doing lots of short sessions of anything.

[Tom]

In this case, though, there is actually a grain of truth to it. Gathering is less effective when you have more fatigue, and each hour of gathering increases your fatigue. Thus, if you gather for 2 hours, then rest for 2 hours, you may have better results on your next 2-hour gathering session, simply because you will no longer be at risk of getting the "you were too fatigued to find anything" message.

However, to calculate effectiveness, you have to counter the 2 hours wasted resting against the probability of having found something in those 2 hours.

So, if you take 10 hours of available time, you can have:
3x2 hours of gathering
2x2 hours of resting

giving you 6 hours of finding something. Let's assume that the probability is 100% (it isn't, but leaving out all other factors, for this direct comparison, let's assume it is). That means you will find 6 items.

Now alternative A would be something like:
4 hours gathering
2 hours resting
4 hours gathering

giving you 8 hours of finding something. Let's assume that the probability goes down 5% every unrested hour after the first 2 (it doesn't, we're just playing "what if" here), then you have:
100%
100%
95%
90%

and that twice. Add that up and mathematically you will find 3.85 items, twice, for a total of 7.7 items. Even if resting for 2 hours only restores half the fatigue (it probably restores more), you'd still get 7.35 items.

Now if you went "!@#$ all this resting" and spent all 10 hours gathering, you'd end up with 8.2 items.


So, unless you assume that the drop is considerably higher than 5%, the advise is still a fail. If it is 10%, then you get 7.4 items in the 2nd and still 6.4 items in the second case.

Basically: Don't trust things just because you've read them on the Internet. :-)


Also: Do trust the game designers when they say the game is written to not reward silly, gamey behaviour.

[Tom]

If you want to prove me wrong, then gather some data. Record each time you visit the academy, including your skill, which trainer, and how many hours (and, of course, whether you gain skill!).

He can't prove you wrong. The game only tells him if he increased his skill, but not by how much. Yes, you can gain more than 1% at once.

[Anaris]

He can't prove you wrong. The game only tells him if he increased his skill, but not by how much. Yes, you can gain more than 1% at once.

Ooh, good point.

[Penchant]

Huh? Randomness affected by skill?

The randomness is random. You could end up with one really good session on the 12-hour one, or a really bad one. Similarly, you could end up with three good 4-hour sessions, three bad ones, or some of each. Or all of the above could end up neutral.

It will average out in the long run. You don't get the better of randomness by doing lots of short sessions of anything.

That was said poorly on my part but it has seemed to me more often that the randomness works out, where like two out of my three smaller sessions are good but when I do the larger one it is 50% for good, so out of 24 hours I would have a very effective 12 and a not so effective one so about half my full potential, where I am getting two thirds my full potential. It obviously is not quite like that nor those numbers but I think you get idea. If I was always lucky the 12 would be better but if its not game mechanic based that the shorter times are better it sure feels like I am just luckier that way.

[Tom]

Yes, it FEELS.

Never forget that human intuition is usually misleading when it comes to statistics and probabilities. Our minds evolved to distinguish patterns, not randomness. That's why we see patterns where there are none - in clouds, for example.

Here's the thing with dice rolls:

If you roll 10 ten-sided dice, and compare it to one 100-sided die, the probability distribution is quite different. With the 10 dice, the probability of getting a very low or very high value are very, very small (the probability to get a 100 is 0.00000001%). With the one die, the probabilities are equally distributed, 100 is as likely as 60 or 50 or 1.

So doing the same randomness-based thing several times instead of once means your end result will be closer to the average, and neither better nor worse. Statistically speaking, if you do it very, very often. On every individual case, it can still be all over the place.


However, that is all only partially true for BM actions. Very few of them are purely random. Usually, they depend on your skill, some region statistics, some other values, and then there's some randomness added. Since these values could change when you do something, doing it again is not simply rolling another die. Depending on how the relationship is, it might be better or worse. For example, if you do something that, say, raises morale, and what you do is more effective with low morale, then doing it once with a large amount of hours is better than doing it in several small steps. But if it is more effective with higher morale, it is the other way around. And when it depends on other values as well...

Basically, stop trying to game the game. There are very few cases where I can imagine that such an approach would yield any measurable advantage. I am very sure that almost all of the cases where something gamey is going around as good advise are purely thanks to some kind of placebo effect.

[Shizzle]

Interesting read - glady I don't usually follow my own 'advice' (resting/gathering/resting is so cumbersome). 16 hour gathering shifts it is!

[Chenier]

I always suspected I was better off with a 16 (or 12) hour run diplomatic act, to minimize losses to what I assumed was rounding down.

[Eldargard]

If I train at the academy for six hours but do not raise my skill, does it make it more likely that I will raise my skill the next time? Of is it simply a random factor each time I try? I assume that my current skill level, the skill level of the trainer, and the number of hours spent all contribute but was wondering if past efforts contribute as well.

[Tom]

It depends.

Basically, the game calculates something like fractional skill levels, so even if you receive a "not enough to improve your skill" message (or no message for the many other places in the game that train skills as a side-effect), mostly that progress is not lost, it is still recorded.

However, it is well possible to actually gain no experience whatsoever. At the academy, it is unlikely, but not impossible.

[Eldargard]

Makes perfect sense. It hurts spending 20+ hours training with no improvement. I feel much better knowing that small steps are being made!

[Tom]

Mostly, you can trust the game messages. If it says you learnt something, but not enough to improve your skill, it means you DID learn something.