Meeting with Maddoc and Absor at the Fan Faire

[Lothay]

...is there a maximum amount of combines before a skillup? basically if i make 1000 combines (regardless fail or success) would i then get a skillup or i it all RNG?

No.

It's all at the whim of the RNG.

[Centuri]

I have had the 300+ combines without a skillup runs. (most notable in my brain was one with Minotaur Hero's Brew before the UI)

However...

I have also had a smithing run (post 200) with 4 skillups in a row.
So, I figure the karma all balances out at some point.

[Phantron]

Even at the lowest % to succeed (2.5%), the chance of failing to skill up after 1000 in a row is roughly 1.0e-11 assuming you've 100% on the first pass. Highly improbable does not imply impossibility, but you should consider yourself very unlucky if this happened to you.

Incidentally, if someone is crazy enough to do a difficulty 4 TS at 70 prime stat, the chance of going 1000 without a skill up would be 1%, so even then it's pretty unlikely (though quite a bit more probable).

[Qaladar Bragollach]

1% assuming what % chance of success? 95%? Wonder what it would look like at 5%.

[Tanker]

Ok, I may have to bring out my Math hat here, as I did a run at computing "streak"-odds before Fan Faire. No one brought up a streak during the panel, so I didn't get to break it out.

Now, streak-odds are a little funny, in that it's the odds that something doesn't happen during a streak, given the odds each time that it does. From my old stat class, given a per-event odds of success of T, the odds that you go N attempts with no successes is: (( 1-T ) ^ N).

As an example, let's use dice. Let's consider a tradeskill success like getting a 1 on a d6. The odds of getting a 1 on a single roll of a d6 is 1 in 6. Now, to compute the odds of getting no 1s in 10 rolls, you really need the odds that you didn't get a 1. Now the odds of not getting a 1 on a d6 roll is 5/6 (or 1 - 1/6). The odds of that happening 10 times in a row is (5/6)^10.

Now let's look at Phantron's example. With a 2.5% chance of success, going 10, 25, 40, 100, 200, 1000 tries without a success looks like:
For 10: (( 1 - 0.025) ^ 10 ) = 77.6% (pretty **** common)
For 25: (( 1 - 0.025) ^ 25 ) = 53.6% (every other person will see this)
For 40: (( 1 - 0.025) ^ 40 ) = 36.3% (even though you expect a skillup ever 40 at 2.5%, 36% of the time it'll take longer to get one).
For 100: (( 1 - 0.025 ) ^ 100 ) = 7.95% ( 1 in 12 people will take at least 100 tries at 2.5%)
For 200: (( 1 - 0.025 ) ^ 200 ) = 0.63% ( 1 in 160 people, this could still easily happen to you )
For 1000: (( 1 - 0.025 ) ^ 1000 ) = 1.01e-09 % (note that this coincides with Phantron's number, converted to a percent) - hopefully only 1 person ever has this happen.

Note that these percentages are the odds of a streak of at least that length occurring. That's why the numbers don't total 100%.

Now, 2.5% is NOT the lowest odds of skillup a person can have, it's only the lowest you can have at pass2. A Pass1 chance of 80% lowers that 2.5% to 2%, which significantly changes the odds above. For example, a streak of 40 at 2% has odds of 44.5%, 8% more overall, but a 22% increase in the odds.

Also note that when I say 1 in 160 people, I do so instead of saying 1 in 160 skillups primarily because you're odds change every time you skillup, so the table above is ONLY useful if your odds are 2.5%. If a success/fail at the combine changes your odds, this table can still be useful if 2.5% is your "combined" chance to skillup, but may not be exactly correct. My stat kung fu isn't strong enough to deal with odds that flip/flop from 1.4% to 2.8% in the same population.

*takes off Math hat*

hope that helps someone out there. You should see Absor's face when I talk like that.

[edits] English goooooood!

[Phantron]

If you've a 70 at prime stat on a skill level 4, your chance of first pass is 70/4 = 17.5%, multiplied by the 2.5% at pass 2, which is about 0.4% total (1 in 250), and with those odds going 1000 without fail isn't very unlikely.

Also note that these are the odds of having X fails on just X combines. If you're doing a total of Y combines, the chance of failing is X in a row is (1-chance of fail X in a row)^(Y-X+1). In simple terms, if the chance to fail 100 in a row is 1%, but you're doing 200 combines, the chance of not failing 100 in a row is the chance to not fail 100 in a row repeated 101 times (combine 1-100, 2-101, 3-102, etc). The odds of this happening is 36%, so it's actually quite possible that you will fail 100 in a row at 1% chance (since you're essentially doing a 1% event 100 times).

I remember reading a book on history of universe and probabilities, and it was saying that when you think how long the universe has been around, the chance of something extremely unlikely to happen is very likely given an extreme amount of time. A planet altering asteroid is believed to hit with a probability of 1 in millions of years, which sounds like pretty good odds, until you factor in the fact that Earth has been around for billions of years. Lottery sounds like pretty low odds (1 in millions), until you factor in that millions of people buy lottery. As reality confirms our hypothesis, lottery do get win by someone with a pretty high frequency (it's probably not you, but someone does win). If you do enough combines, eventually some very lucky or unlucky event is bound to find you.

[Dhomal the Scavenger]

Hello-

Quote:
"Another topic we discussed was fletching. Absor flatly stated that fletching was messed up as far as skilling up. This is something the devs know about and they're actively working to get some alternatives out there"

I have a simple question about this (*I hope!*)

What is the problem? Is it lack of skillup choices? Is it something else that causes problems in skillups?

I've been merchant-diving for Acrylia and condensed for Acrylia Arrows - and am slowly (*read, very slowly!*) inching up towards 268 with them. I had planned on possibly working on Nightmarewood bows afterwards - but I notice that the Nightmare Arachnid Silks seem to be needed for some of the DoN cultural stuff.

Just had not seen any additional mentions of this aspect - so I thought I would ask.

Dhomal the Scavenger
Warrior 65 - Cazic Thule
Brellian Defenders

"Questions are not inherintly stupid. Just sometimes mis-informed."

[Sylphan]

Acrylia arrows are great to 268, and there's even some recipes up to 282.

The problem comes when you start doing Nightmarewood bows. You'll find it's much much harder to get the parts Nightmarewood than it was for acrylia arrows. Getting to 300 becomes impractical. But other people have cultural fletching, and they find it very easy to get to 300.

[KyrosKrane]

The problem with fletching happens (kinda) from 269-282 and particularly from 283-300. In that latter range, there's something like a dozen recipes (with a few minor variations of some recipes, like multiple versions of arrows or bows) you can do to skill up. The drop rate of the rare bits to make those recipes is atrociously low. (This is stuff like Bazu bows, DoN and some Omens arrows, and PoP bows.) You simply can't collect a sufficient number of rare bits to get enough combines to skill up in that range. Part of the solution is to look at improving the drop rates or adding more recipes; another part is the enhanced chance to skill up on these combines. The devs said they would look at some combination of these two parts as the solution for fletching, but they didn't say when that would be done.

[Verdandi]

Any objections to unstickying this? Kyros?

[KyrosKrane]

I think most of the issues in my post have now been addressed, and the outstanding stuff is either well-known or has its own post elsewhere. Unstickying this should be fine.

[Verdandi]

Unstuck it is. If it's needed it'll still be around.