Barrr noncombat stat gains

This is where you spade, shovel and sickle. Or is it reap?
Post Reply
User avatar
QuantumNightmare
AFH
Posts: 1827
Joined: Mon Jan 29, 2007 10:28 pm
Location: Montreal

Barrr noncombat stat gains

Post by QuantumNightmare » Sat Mar 08, 2008 12:49 am

Kujjie ran some turns, I wrote a parser. A bit of math later, an equation pops up.

Testarrrsterone (wuss out choice) : gain moxie equal to 2*mainstat, capped at 150
Rock Starrr (low places choice with 19 drunk): 1*mainstat to all stats, capped at 100

Image

Image

Seal Lubber
AFH
Posts: 318
Joined: Fri Sep 21, 2007 12:27 am

Post by Seal Lubber » Sat Apr 05, 2008 10:10 pm

(I'm no maths guy, so please be critical)

Rock Starr - Option 3
-- Mainstat 14, Drunkenness 0-8
-- Gives all three stats, in the range of (Mainstat/2-Mainstat) to each, uniform distribution. Summarised below:

Code: Select all

Gain #
7    15
8    15
9    16
10   15
11   13
12   17
13   9
14   20
     120 (40 non-combats, 3 data-points per)

-- Main stat gain would be (M+M/2)/2
-- Total stat gain would be 3*((M+M/2)/2)
-- I'll try testing at different mainstat. It seems to contradict the 1*mainstat found above. EDIT: Scrap that, I'm getting much bigger stat gains with 10 drunk, what I've found here is the 0-9 drunk stat gains, with QN finding the 10-19 gains above.
-- Drunk level had no effect that I could tell, data is available if anyone wants a look.

Testarrrsterone-Option 1
-- Mainstat 14, drunkenness 0-8
-- Gives mus/mox stats, in the range of (Mainstat/2 to Mainstat) to each, uniform distribution. Summarised below:

Code: Select all

Gain #
7    9
8    11
9    5
10   8
11   12
12   8
13   12
14   9
     74 (37 non-combats, 2 data-points per)

-- Main stat gain would be (M+M/2)/2, for moxie and muscle
-- No myst gains, since I'm not myst class (or whatever the requirement is). EDIT: I just dropped my muscle, so my stats are 41 51 15, but I still don't get the myst gains. The wiki says you'll get myst stats if Base Myst is highest stat, which is untrue.
-- Does this need testing at different mainstat?
-- Drunk level had no effect that I could tell, data is available if anyone wants a look..

User avatar
QuantumNightmare
AFH
Posts: 1827
Joined: Mon Jan 29, 2007 10:28 pm
Location: Montreal

Post by QuantumNightmare » Sun Apr 06, 2008 12:48 am

Nice clean results. I wonder why I never thought of testing statgains like this... keeping mainstat constant and seeing what the range would be.

Two reasons to vary mainstat: find out what happens near the upper end of the scale where statgains hit the max, and to find out if the variation you see is based on a % of mainstat, compared to a set variation.

Either way we've got two new formulas, and an easy way to find the rest.

Seal Lubber
AFH
Posts: 318
Joined: Fri Sep 21, 2007 12:27 am

Post by Seal Lubber » Sun Apr 06, 2008 3:11 am

QuantumNightmare wrote:Two reasons to vary mainstat: find out what happens near the upper end of the scale where statgains hit the max, and to find out if the variation you see is based on a % of mainstat, compared to a set variation.
I did the tests at mainstat 14 because it's convenient to do with steroids which give -15 stats, so I could maintain my mainstat by using a single steroid whenever my moxie went up a level. Going higher is much more expensive and less convenient. I'll see what I can do though.

EDIT:
Rock Starr - Option 3
-- Mainstat 19, Drunkenness 0
-- Just confirming for different mainstat: Gains of M-(M/2)

Code: Select all

Gain #
9    4
10   5
11   5
12   3
13   3
14   4
15   7
16   3
17   5
18   1
19   5
    45 (15 noncombats)

-- Limited data, but: Since the number of 9's (4) is close enough to the average (4.09), this would suggest that the M/2 gets floored when determining the range. Is this typical KOL behaviour, or would you expect something else?
-- The problem with testing around the upper end of the scale is that it costs a lot to delevel, and the larger spread of gains will mean many tests might be needed to confirm the range. You'd think it would just roll a number from the normal range, with any results higher than the max being rounded down; is this easy/hard to confirm?

Testarrrsterone-Option 3 (Wuss)
-- Mainstat 19, Drunkenness 0
-- The formula in QN's first post is 2*mainstat, I'm getting 2.5*mainstat:

Code: Select all

Main Gain
19   56
19   44
19   49
19   40
19   56
19   50
19   54
19   39
19   49
19   54
19   41
19   53
19   41
19   51
19   53
Avg  48.7

-- This gives a range of 39-56, or (2.05*mainstat)-(2.94*mainstat). Close enough to 2-3*mainstat.
-- Lack of data in the original post? Related to my lower mainstat?
-- As I keep saying, this does start to get expensive with higher stats, so let me know what you think would be a good test so I don't waste too much meat/time.

Seal Lubber
AFH
Posts: 318
Joined: Fri Sep 21, 2007 12:27 am

Post by Seal Lubber » Wed Apr 09, 2008 9:55 am

Testarrrsterone-Option 3 (Wuss)
A chance to try out Excel 2007 (I still can't find a fucking thing in those menus):
- First graph is data collected over the last few days showing 2-3*mainstat gains with a cap of 150.
- Second graph shows some data that QN sent me, I'm not sure what the deal is here.

Image
Image

QuantumNightmare wrote:...find out what happens near the upper end of the scale where statgains hit the max.
8 tests at 60 mainstat (giving a mean gain of 150) gave four gains of 150 and four gains of <150. This is what you would expect if the distribution remained the same with values higher than the cap being rounded down.

-- Are you guys interested in the various drunk giving stat adventures? I'd probably ascend the multi for another cyrpt run and hit these up next time if you're interested. If you want it done now, this can also be arranged.
Last edited by Seal Lubber on Fri Apr 18, 2008 2:22 pm, edited 1 time in total.

User avatar
KujjieKujjieKoo
Spade Ninja
Posts: 2240
Joined: Mon Apr 30, 2007 8:50 pm
Location: !Chicago
Contact:

Post by KujjieKujjieKoo » Wed Apr 09, 2008 10:07 am

I've got a variety of debug logs of barrr spading data, if you want it Seal. Just tell me where to send the logs!

User avatar
QuantumNightmare
AFH
Posts: 1827
Joined: Mon Jan 29, 2007 10:28 pm
Location: Montreal

Post by QuantumNightmare » Wed Apr 09, 2008 12:24 pm

Kujjie, I parsed all the logs you sent me and sent them all to Seal Lubber in an excel sheet. He's already looking at your results, at least the one's I've seen.

That second graph doesn't make any sense... it's one reason why I stopped looking at this. No time to figure out WTF is going on... since the analysis doesn't look to be simple. Something else is effecting stat gains.

User avatar
Flolle
Fie the Pie
Posts: 848
Joined: Mon Apr 02, 2007 10:13 am

Post by Flolle » Mon Jul 07, 2008 7:57 pm

Is someone still working on this? I was just curious. (probably not, what with the run on Hobopolis and all)

Also, I think it's interesting that the stat gains from the noncombats in the barrr use another scaling mechanism than most of the other scaling noncombats (at least the ones I can think of). Here, we have a hard cap which is first reched by the upper end of the scaling and quite a bit later by the lower end, which is quite different from what we have for example at the castle. The result is that the average stat gain will need much more time to reach it's peak. Imagine a line between the two borders Seal Lubber used in his last post's images to mark the upper and lower end of the scaling. Note that once the upper scaling reaches the cap, the line between the two borders will climb only half as fast as it did before.

I never really noticed that before, which might have been just me. But my question is: Is the barrr the only place that acts like this, or are there others?

User avatar
QuantumNightmare
AFH
Posts: 1827
Joined: Mon Jan 29, 2007 10:28 pm
Location: Montreal

Post by QuantumNightmare » Mon Jul 07, 2008 8:37 pm

No one is working on this right now. We should be, but this is one of a long list of orphaned projects.

Library noncombats also work in this way, but most of the main leveling zones don't. Not that it makes much of a difference.

Post Reply