Optimal place to level (Updated 2/05/09)

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

Optimal place to level (Updated 2/05/09)

Post by QuantumNightmare » Sat Mar 24, 2007 9:05 pm

Here's the link: http://sites.google.com/site/quantumnightmare/Home/nu-stat.xls?attredirects=0

FAQ:

Q. How are PLD stats determined?
A. PLD stats are calculated for a zone that has been set up for powerleveling. The sheet assumes someone has already done unlimited cold water diverts, and that the club will never close as you continue dancing for stats.

Q. Why is the gallery combat rate so much higher than it should be?
A. I consider knight noncombats to be combats, so the combat rate is the odds of getting anything that isn't a louvre.

Q. How are Ed stats calculated?
A. By assuming you run away on the 7th form continuously, losing an adventure each time.

Q. Why are my absinthe gains so much lower than what the sheet predicts?
A. The sheet calculates absinthe gains when using the 9-5-1 method of adventuring for as few turns as possible while still getting all three noncombats.

Q. Why do the ballroom calculations never change?
A. The ballroom uses plenty of circular references, and your spreadsheet software needs to have iterations enabled in order to be able to calculate these circular references. To do so in excel 2003, go to Tools -> Options -> Calculations... and then make sure the "iteration" option is checked. To do so in Office for OSX, go to preferences -> calculations. I don't think OpenOffice can do iterations, but haven't tried myself.

Q. The stat day option is broken. I type "Yes", and nothing changes.
A. The correct input for stat day is "muscle", "myst", or "moxie". While the default setting is "no", you must tell the sheet which stat day it is for it to calculate anything.

Q. What is the moxie setting used for? It's blank on my sheet.
A. Wormwood combat stats scale to your moxie. As a moxie class, your mainstat is used to calculate the power of these scaling monsters, but as a myst or muscle class an approximation has been made. While a relatively large difference in moxie values will make a relatively small difference in your stat gains, the moxie setting can be used to give a more accurate reading to non-moxie classes adventuring in the wormwood.

Q. <Your next question here>
A. Feel free to post any questions or comments in this thread, as I'd like this tool to be as intuitive, useful and accurate as possible.

---------------------------------------------------

There is no "best" place to level. It depends on too many factors for a guide to tell you where to go. Gut feeling works pretty well, but I wanted something more concrete. So I threw together a spreadsheet that compares some of the top leveling locations.

On the top left, you input some stats about your character. Your mainstat type, familiar weight and type, +ML value, and combat modifiers are the most important. But everything that can effect stat gain should be included. Then in bold, you pick what your mainstat value is, and what stat you are trying to get. The sheet calculates everything else, and plots out the expected return from the gallery, ballroom, palindrome, castle, battlefield, and bedroom.

Once again, the only values you touch are in the top left box, under "User-defined variables". And the top two in bold are the only things you should touch more than once. Then just read across the top, to find the best place to adventure. All the numbers beneath are just calculations. Pay no attention to that man behind the curtain!
Last edited by QuantumNightmare on Thu Feb 05, 2009 4:07 pm, edited 36 times in total.

blackbelt02

Post by blackbelt02 » Fri Mar 30, 2007 8:33 pm

I assumed the best place to level after having 68 of the main stat was the gallery. Am I incorrect?

Raccoon
Oh my! Guy with Pie!
Posts: 1975
Joined: Thu Feb 01, 2007 10:33 am
Location: somewhere on the West Coast

Post by Raccoon » Fri Mar 30, 2007 8:37 pm

blackbelt02 wrote:I assumed the best place to level after having 68 of the main stat was the gallery. Am I incorrect?


Others seem to like the bedroom or ballroom, for myst or moxie classes. But in HC, I stick with the Gallery because I don't need to mess with opening the second floor, etc. Plus, Gallery provides wads and, for myst classes, possible food.
[img]http://i169.photobucket.com/albums/u218/mad_hamish/raccoonsig.jpg[/img]

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

Post by QuantumNightmare » Sun Apr 08, 2007 12:47 pm

Can this post be moved to the commune? I'll use this post purely for updating my excel sheet, and the main area comparison for regular discussion.

Sinshlad
Inscrutable Pi
Posts: 273
Joined: Tue Jan 30, 2007 7:40 am
Location: Australia

Post by Sinshlad » Sun Apr 08, 2007 1:01 pm

Ha! finally a reason to use my mod powers for something other than deleting posts made by an idiot.

Anyway, I plan on making use of this. It should save me time that I would normally be spending doing math mid-run.
[img]http://img.photobucket.com/albums/v300/dvmaprezson/AFHV1.gif[/img]

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

Post by QuantumNightmare » Sun Apr 08, 2007 2:09 pm

My thoughts on the ballroom are dropping quickly. Once I factored in the rejection of even a small amount of quartet songs and the adventures taken from using dance cards, the palindrome almost always beats the ballroom for moxie. I can't see a reason to go to the ballroom anymore except to open the song. The sheet was updated once again.

I also threw in another interesting thing on sheet #3 of that same excel sheet. Bordered by bold, you set the table size you want. When powerleveling, set it to start at your mainstat, and end at the desired value. Say, from 104 to 200. Then fill in how many adventures you have left every so often when you gain a statpoint, and it'll calculate how many stats/turn you need to reach the desired goal with the adventures you have left. It helps make sure I'm on the right pace!

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

Post by QuantumNightmare » Wed May 30, 2007 8:24 pm

I've updated this again, and cleaned it up a bit. The calculation page is now separated from the results, and the bathroom is included. The main issue now is my handling of the bedroom. I assume you take all stat adventures, bit I'll need to work out the math for if you only want a single stat.

Everything else should be included - quartet odds, dance cards, castle wheels....

Mercenary9
Inscrutable Pi
Posts: 259
Joined: Mon May 21, 2007 12:32 am
Location: California
Contact:

Post by Mercenary9 » Wed May 30, 2007 9:16 pm

QuantumNightmare wrote:My thoughts on the ballroom are dropping quickly. Once I factored in the rejection of even a small amount of quartet songs and the adventures taken from using dance cards, the palindrome almost always beats the ballroom for moxie. I can't see a reason to go to the ballroom anymore except to open the song. The sheet was updated once again.


So it's been proven that this song works everywhere and not just the ballroom?

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

Post by QuantumNightmare » Wed May 30, 2007 9:18 pm

yes. Choices are +5ML, +5% noncombats, +5% items.

theophrastus
Spy vs. Pie
Posts: 77
Joined: Thu Apr 19, 2007 5:45 am

Post by theophrastus » Thu May 31, 2007 8:19 am

Thanks, QM, pretty interesting. I am going for a L15 skill, so this is really useful. I'd thought the +5 ML song would be better when using a sombrero, but apparently the old "--combat is best" theory holds here. According to the spreadsheet, though, the ballroom is still better for moxie for me by about 3 moxie substats per adventure. That seems pretty substantial; I wonder why your results are different.

-theo

VladimirPootin
AFH
Posts: 1236
Joined: Tue Mar 27, 2007 2:03 am
Location: Ad Absurdum

Post by VladimirPootin » Thu May 31, 2007 12:43 pm

mine and QN's? i don't totally trust my ballroom numbers, but remember they depend a lot on drop rate.
[b]AFH: Now with more [i]new and improved[/i] moral ambiguity![/b]

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

Post by QuantumNightmare » Thu May 31, 2007 6:38 pm

I used an approximation for the quartet odds, capped at 4.5%. It's not exact, but it's accurate enough. From that point on, everything should be good in my ballroom handling.

VladimirPootin
AFH
Posts: 1236
Joined: Tue Mar 27, 2007 2:03 am
Location: Ad Absurdum

Post by VladimirPootin » Thu May 31, 2007 9:49 pm

i used an even 50% for the quartet... thinking about it i forgot the spading. need to fix up the numbers.
[b]AFH: Now with more [i]new and improved[/i] moral ambiguity![/b]

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

Post by QuantumNightmare » Fri Jul 13, 2007 3:23 pm

This sheet has been updated for the new volleyball equation, and ML stat gain. Sombrero and fairy need to be spaded, as well as the effect of +ML on combat percentage. Otherwise, it's accurate.

lordhades15
Fie the Pie
Posts: 1175
Joined: Sat Feb 03, 2007 12:02 am
Location: Kent, OH

Post by lordhades15 » Fri Jul 13, 2007 4:37 pm

Keep in mind the palindrome is now unavailable until level 11. So, it isn't really a leveling consideration unless you are powerleveling (I suppose it is possible that it could have an impact bridging the gap from 11 to 12, but meh).
j12601 wrote: Turns are now sub-optimal. It is a brand new fucking day.

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

Post by QuantumNightmare » Fri Jul 13, 2007 7:19 pm

Ah, ok. I'll get rid of that column until you hit level 11. Actually, I might get rid of it altogether, until I understand what is going on with the new noncombats.

User avatar
Mad Hamish
AFH
Posts: 1517
Joined: Mon Apr 09, 2007 8:03 pm

Post by Mad Hamish » Fri Jul 13, 2007 7:47 pm

QuantumNightmare wrote:Ah, ok. I'll get rid of that column until you hit level 11. Actually, I might get rid of it altogether, until I understand what is going on with the new noncombats.
They've seemed to all be superlikelies once you have the quest, but the shelf adventure wont show until you have all 4 items. But this is a very small sample size. Me, my multis, and my girlfriend.
Image
Hi-keebated again? No way!

User avatar
Majin Vlad
Inscrutable Pi
Posts: 311
Joined: Thu Jun 21, 2007 10:54 am
Location: Brooklyn, New York
Contact:

Post by Majin Vlad » Fri Jul 13, 2007 8:59 pm

Mad Hamish wrote:
QuantumNightmare wrote:Ah, ok. I'll get rid of that column until you hit level 11. Actually, I might get rid of it altogether, until I understand what is going on with the new noncombats.
They've seemed to all be superlikelies once you have the quest, but the shelf adventure wont show until you have all 4 items. But this is a very small sample size. Me, my multis, and my girlfriend.


Myself,my multi and a few of my friends have also seen that the shelf doesn't show until you have all 4, so I can pretty much confirm it.

Once you hit level 11 there isn't really a need to level up anymore...after that everything should be going pretty smoothly.
[img]http://img8.imageshack.us/img8/8913/afhk.jpg[/img]
[url=http://www.giantitp.com/Comics.html]Order of the Stick Comic[/url]

lordhades15
Fie the Pie
Posts: 1175
Joined: Sat Feb 03, 2007 12:02 am
Location: Kent, OH

Post by lordhades15 » Fri Jul 13, 2007 9:31 pm

I think that the palindome quest items are actually noncombats.
Two reasons:

a) the palindome has an already high noncombat rate, which can make things seem superlikely,

b) I got all three quest items in a row with noncombats cranked this run. I pulled the ketchup hound and spent 1 more adventure in the dome, and guess what? Got the shelves. So unless I was really freakin lucky and just happened to get all four in a row, noncombats seem likely.
j12601 wrote: Turns are now sub-optimal. It is a brand new fucking day.

User avatar
Mad Hamish
AFH
Posts: 1517
Joined: Mon Apr 09, 2007 8:03 pm

Post by Mad Hamish » Sat Jul 14, 2007 12:15 am

lordhades15 wrote:I think that the palindome quest items are actually noncombats.
Maybe I'm wrong, but your data sounds more consistent with superlikelies than regular noncombats. I've also seen the 3 quest items show up within 4 adventures every time the MMM&MG spading triumvirate (see last post) have gone through this quest. The lady doesn't boost non-combats generally, nor did Ninja is a Verb... Those adventures show up REAL easy.
Image
Hi-keebated again? No way!

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

Post by QuantumNightmare » Sun Jul 15, 2007 9:44 pm

I've updated the sheet again. Two updates this time: palindome is hidden before level 11, and the bedroom calculations are now set to optimize a single stat. Combats are chosen over offstat noncombats if given the choice.

VladimirPootin
AFH
Posts: 1236
Joined: Tue Mar 27, 2007 2:03 am
Location: Ad Absurdum

Post by VladimirPootin » Mon Jul 30, 2007 1:49 pm

maybe everyone already knew this, but the palindome now gives a lot more stats with papayas... i'm getting 300 of each stat with papayas, and 100 of each without.
[b]AFH: Now with more [i]new and improved[/i] moral ambiguity![/b]

Brulak
AFH
Posts: 945
Joined: Sat Apr 14, 2007 2:51 pm
Location: Hamburg, Germany

Post by Brulak » Mon Jul 30, 2007 3:06 pm

Already was that way before NS-13, however, you speed folks never leveled beyond 11. Until then (and in Ronin), it was cheaper to collect the 100 per stat than go through getting papayas first.

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

Post by QuantumNightmare » Mon Jul 30, 2007 8:26 pm

As brulak says, this the palindome was always like that. My sheet assumes that at 200 mainstat you switch to getting papayas/trading them in, instead of taking the pep talk. Of course, if you tell it you have papayas it will always trade them in, and you get even more stats.

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

Post by QuantumNightmare » Tue Aug 14, 2007 11:43 pm

A few NS-13 updates:

-Odor extractor in the ballroom is now an option
-Sombrero formula added
-Random guess fairy formula as well (to be taken with at least 3 grains of salt)

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

Post by Flolle » Wed Aug 15, 2007 6:44 am

Yay, new update to my favorite tool to check whether I screwed up! ;)

Question: Is it possible that you are calculating the fairy bonus from the second page always into the ballroom calculation even when using a nonfairy familiar?

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

Post by QuantumNightmare » Wed Aug 15, 2007 6:14 pm

Nope, although it does seem that way. The "fairy bonus" is a very rough approximation of the effect a fairy at that given weight would have, but it's only used in the calculations if the familiar is actually a fairy (or shaman). I'll make that calculation more clear next time I update this.

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

Post by Flolle » Wed Aug 15, 2007 7:09 pm

That's ... strange. I get totally different output than you (yours is much higher). And I have no idea why.


Edit: Ok, I tinkered a bit around and apparently the statday option is borked. Maybe this screws up other calculations? Also, the modified noncombat rate for the ballroom on the second sheet looks strange to me.

It really gets me agitated that I have no idea if I'm doing something wrong in the ballroom (though I don't see were ... :?). So please don't feel offended.

Also, "Songs appear 15% as often as curtains", were did you pull that from? I thought they had the same appearance rate whenever the quartet can appear. Or am I misunderstanding something?


Edit edit: I updated to OpenOffice 2.2, but still no dice. I'm getting the same different values from the spreadsheet which I told you about in chat. This is very aggravating. :? (But not your fault!)

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

Post by QuantumNightmare » Tue Aug 21, 2007 1:44 pm

Minor updates
-stat day calculation is fixed (has been for a while, just didn't post about it)
-Combat ratio now takes into account that +1ML cancels out approximately +0.2% noncombats. The first +10ML is ignored, assuming that you get it from an MCD source.

As well, some explanation for my numbers. The combat rate in the ballroom looks OK to me. I first calculate the combat rate taking into account combat modifiers and the presence of matilda (where a certain percent of your time spent in the ballroom does not go to either combats or noncombats, but to dance card adventures). Then I take into account skipping the quartet, which increases the combat ratio again.

15% quartet to curtains... while it's true that when the quartet can occur it is just as likely as the curtains (usually much more, due to the queue) I've found a simple way of approximating the overall quartet rate to be 15% of curtains. There is no math behind this, it's just that this value gives a reasonably close approximation of the quartet odds as found by AFH's past spading. And plugging in exact values for each combat modifier possible doesn't effect the stat gain by an appreciable amount. This should be fixed, but it's not worth the effort right now as 15% is good enough.

Finally, if both are sheets are giving different values then something is wrong. Besides the fact openoffice and excel calculate different values (bug in circular calculations), you should still be getting the same values as I do on my PC. That means one of us is making a mistake...no idea who.

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

Post by Flolle » Fri Aug 24, 2007 9:28 pm

OK, I'm writing this as I'm looking through the code of my tool once again:

- the Ballroom has a standard combat frequenzy of 80%
- all combat encounters have equal probability and the ML of all of them is 110
- dance cards have a 15% drop rate
- the Rotting Matilda adventure fires after 3 normal adventures and has a 2*mainstat scaling, capped at 300
- dance cards are used after one another so as to not interfer with that counter
- both noncombats have the same appearance rate
- but the Quartet can only occur once every 20 adventures
- the Quartet can be skipped but gets added to the queue anyway
- the Curtains adventure has choices for either combat or moxie gain (1*mainstat, capped at 150), the choice with better gains for the desired stat is taken
- of course, normal non/combat queue handling is in effect

If this all is right and I have not forgotten something, I have implemented the Ballroom perfectly, as far as current spading permits. Of course, I'm actually rolling the encounters and am thus subject to RNG, but 1000000 rolls should average things out. It is also possible that I've made a fault in the general encounter rolling, but my numbers seem to work for other areas, which would imply that I did not. Other faults are also possible, but I've looked through the code many times and haven't seen anything. That doesn't necessarily mean anything, but the source is free, so if you want to double check, please do so. In fact, I'd be really happy if someone did.

I hope you understand now why this gets me so agitated. If your values are correct, it would mean mine are not, even though I've looked through the code so often and have no idea where there might be an error in my calculations.

As a little example I did a calculation with these parameters. I cannot check what output your sheet will give, because I'm using OpenOffice, but it's more of an example for you to see how big the difference between both our tools are. Maybe you'll get the values I do with MS Office, which would nullify everything I said. That'd be great. :D

Mus substat gain: 6.05
Mys substat gain: 6.05
Mox substat gain: 38.96
Total substat gain: 51.06


P.S.: I have to sincerely apologies for derailing this thread in such a way and coming off as a big jackass. I'm only doing so with good intentions.

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

Post by QuantumNightmare » Sat Aug 25, 2007 2:52 am

With those same parameters, I get:

Muscle: 5.2
Myst: 5.2
Moxie: 44.0

Calculations: 80% base combat, at 110ML. Curtain gives 1*mainstat moxie capped at 150, and is always taken (should be changed, but is good for this case where moxie is taken anyways). Matilda gives 2*mainstat capped at 300. At +15% noncombats, quartet will occur 4.6% of the time and be skipped. This number was previously spaded.

Turns per dance card are guessed randomly. Call this X. Combat rate is found to be (80%-15%) * (X-1)/X. This means 65% combats are found, but if every 13 turns a dance card adventure is found, then the total combat rate is really 65% * 12/13 = 60%.

Now quartet is skipped. Every time the quartet is rolled, you have a 60% chance of getting a combat. The quartet can be skipped, rolled again, skipped, rolled again, skipped....

Combat rate = 60% * (1 + 4.6 + 4.6^2 + 4.6^3 + 4.6^4 + 4.6^5 + 4.6^6 + 4.6^7 + 4.6^8) = 62.9%

Turns per dance card are calculated. Dance cards drop from every third monster with a 15%*250%=37.5% chance. So at 62.9% combats, we get a dance card every 1/(0.629/3*0.375)=13 turns. This number is then used to calculate the combat rate, and the whole calculation is started again. This is the circular calculation, and it's found that a dance card is actually found every 13th turn. Usually, excel does a couple dozen of these until the right number is found.

Now. Total combat stats are [ 110ML/4 + sqrt(30) ] * combat rate = 20.7
Total noncombat stats are [mainstat * noncombat rate] = 75 * [1-62.9%-1/13] = 22.1
Matilda stats are 2*mainstat * 1/13 = 11.58

Total stats:

Muscle = combat stats/4 = 5.175
Myst = combat stats/4 = 5.175
moxie = combat stats/2 + noncombat stats + matilda = 10.35 + 22.1 + 11.58 = 44.03

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

Post by Flolle » Sat Aug 25, 2007 7:48 am

First of all, you're post is really great. I finally understand what you are actually doing in your sheet. This helps immensely. :)

I went through your post and this is what sprang to my attention:
QuantumNightmare wrote:... Dance cards drop from every third monster with a 15%*250%=37.5% chance ...

I did the calculations with +150% to itemdrop (Edit: Which now that I look at this again seems to be the case for you, too. Was a little fast here. :oops: But the fact that my calculations at +250% are close still stands.). Could it be that you are haveing a hidden bonus to itemdrops somewhere that shouldn't be there? Because with +250% itemdrops I'm getting:

Mus substat gain: 5.83
Mys substat gain: 5.83
Mox substat gain: 43.1
Total substat gain: 54.76

This is close enough to your values that I would put it down to the different approach with the calculations in our programs.

Otherwise I couldn't find anything at fault with your math.

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

Post by QuantumNightmare » Sat Aug 25, 2007 10:48 am

Item drop formula: (Base rate) * (1 + item bonus)
With no bonus: 15% * (100% + 0%) = 15%
With +150% bonus: 15% * (100% + 150%) = 37.5%

It looks like this is your mistake, if you multiply 15% * 150% you are only adding a 50% bonus. When you took this into account and added a bonus of "250%", we get pretty much the same numbers. Your ballroom data is actually more exact that mine, since I'm forced to use a few approximations because the ballroom is so complex.

And please, use this thread to point out times where my sheet breaks, or doesn't give the same number as yours. The point is to have the most exact approximations we can have, so it's good when we can point out my bugs (or yours!). As you can see, the math I use is complicated enough even when you know what's going on!

EDIT: New version of my sheet is out, with bathroom stat gain corrected, and bombed out frats added.

Alrik
My Pie Blown Sky High
Posts: 613
Joined: Fri Jun 08, 2007 7:08 pm
Location: near Toronto, Ontario

Post by Alrik » Sat Aug 25, 2007 4:57 pm

Thanks to both of you. The resources you're making available are amazing. =)

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

Post by QuantumNightmare » Tue Aug 28, 2007 7:06 pm

New version released today. A couple new features, mainly to help with powerleveling.

- Meat calculations for the gallery, palindome and castle. Calculates meat drops and autoselling of items (only jumbo olives are mall sold)
- New zone: sign gym
- New option: Buy clovers with meat drops and use them in appropriate zone. Only works for gallery, palindome and castle

Raccoon
Oh my! Guy with Pie!
Posts: 1975
Joined: Thu Feb 01, 2007 10:33 am
Location: somewhere on the West Coast

Post by Raccoon » Tue Aug 28, 2007 8:31 pm

QuantumNightmare wrote:New version released today. A couple new features, mainly to help with powerleveling.

- Meat calculations for the gallery, palindome and castle. Calculates meat drops and autoselling of items (only jumbo olives are mall sold)
- New zone: sign gym
- New option: Buy clovers with meat drops and use them in appropriate zone. Only works for gallery, palindome and castle


Nice -- but of course, the Palindome offers the option of buying papayas (x3), which at high enough level is the same as a clover, but considerably cheaper (anywhere from 1200 to 3000 for 3 papayas).
[img]http://i169.photobucket.com/albums/u218/mad_hamish/raccoonsig.jpg[/img]

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

Post by QuantumNightmare » Tue Aug 28, 2007 8:46 pm

If you mean it might be best to always take papayas, sell them, and buy dance cards / clovers... that depends on the market but I might as well work it out.

Otherwise I already have an option (Got Papayas?) that does calculations based on unlimited papayas. If you don't have unlimited papayas, after a certain point the sheet will tell you to start collecting and trading in papayas instead of taking the pep-talk. If you meant something different, let me know.

Also, you now buy dance cards for moxie, clovers for muscle and myst.

User avatar
Prestige
AFH
Posts: 314
Joined: Mon Apr 16, 2007 8:46 pm

Post by Prestige » Sun Sep 02, 2007 8:29 pm

I was wondering about the usefulness of absinthe-aided leveling. As you know, with each 10 turn shot of absinthe, you can guarantee 3 adventures that give you 1*(mainstat) of the stat of your choice, capped at 150. You will also have to spend between 0 and 3 turns (on average 2) fighting worm wood monsters. The remaining turns can be spent in the zone of your choice.

So what do you think? My rough calculations say that it's better than (or rather "good in addition to") normal adventuring in almost every situation.

snarles2
Spy vs. Pie
Posts: 81
Joined: Tue May 15, 2007 4:12 pm
Location: BALRAWG!!

Post by snarles2 » Sun Sep 02, 2007 8:32 pm

Last two, yes, first one, not so much.

It's worth it to harvest 3-4 bottles a day for this.

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

Post by QuantumNightmare » Sun Sep 02, 2007 11:10 pm

You asked, so I added an absinthe zone. Some assumptions that went into the stat calculations:

- Monsters have ML = moxie, capped at 125. You now have the option to enter your moxie at the bottom left of the page, and if this is left blank your moxie will be approximated from your mainstat. If you're a moxie class, it takeyour full mainstat, otherwise it assumes moxie = 70% of your mainstat.

- Wormwood can be completed in 3 to 8 turns, average of 5.5. Out of this, 2.5 will be combats. This gives a combat rate of 45.5%, with 3 stat noncombats out of 5.5 turns.

Note, this is only for the combats in the absinthe zone itself. The ~4.5 turns you can skip are best spent in the next best leveling location.

Also, you might get a warning about macros in this version. Just disable them, I was fooling around with writing macros to draw graphs from my data and tried to delete them. Nothing bad will happen if you enable the macros, but nothing useful will happen either.

User avatar
Prestige
AFH
Posts: 314
Joined: Mon Apr 16, 2007 8:46 pm

Post by Prestige » Mon Sep 03, 2007 1:34 am

Interesting work, QN. Thanks for adding those calculations.

QuantumNightmare wrote:- Wormwood can be completed in 3 to 8 turns, average of 5.5. Out of this, 2.5 will be combats. This gives a combat rate of 45.5%, with 3 stat noncombats out of 5.5 turns.


In my experience though, completing the Worm Wood (getting all 3 stat gains) only takes 3 to 6 turns, 4.5 on average. The first one takes 1-3 turns, the second 1-2 and the last only 1.

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

Post by QuantumNightmare » Mon Sep 03, 2007 9:17 am

The wiki writes that the first noncombat appears on turns 9-6, the second on turns 5-3, and the third can be forced on turn 1. That's 8 turns total, 3 turns minimum. I seem to remember that it can be done in less, so if anyone can show me some numbers I'll adjust the stat gains in the absinthe area accordingly.

feng shweez
Inscrutable Pi
Posts: 361
Joined: Wed Jul 11, 2007 10:41 am
Location: Boston, MA

Post by feng shweez » Mon Sep 03, 2007 1:16 pm

The first one is always 9-7 I believe. The second one is always 5-4, I thought.

I could be wrong though.
[url=http://img341.imageshack.us/img341/9392/nootrosforumslk9.gif][img]http://img341.imageshack.us/img341/9392/nootrosforumslk9.gif[/img][/url]

dalgar
Pie of the Sky
Posts: 26
Joined: Sun Sep 02, 2007 9:32 am

Post by dalgar » Mon Sep 03, 2007 2:03 pm

Feng is correct on both counts. This has been spaded on the HCO forum, to which I cannot provide a link because I have a small epeen (postcount). Heh. It's in Waiting4Oct's Not-a-pipe Yields thread.

Greetings, by the way.

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

Post by QuantumNightmare » Mon Sep 03, 2007 4:01 pm

9-7 might be right, but just today I got an absinthe noncombat with 3 turns remaining. Granted, I was adventuring for 7 turns in a row in the wormwood and then skipped to 1 turn remaining for the last noncombat. I'll edit my sheet to include 9-7, 5-4, 1. This does make absinthe an even better leveling zone. Good call, prestige.

Was my noncombat at turn 3 an oddity that will never happen if i had skipped turn 10 and turn 6? Also, welcome Dalgar.

User avatar
Prestige
AFH
Posts: 314
Joined: Mon Apr 16, 2007 8:46 pm

Post by Prestige » Mon Sep 03, 2007 5:18 pm

I've done 9-7, 5-4 and 1 many times and it's always worked. I am curious though as to what happens if you skip parts of this pattern. Maybe I'll do some experimenting tomorrow. Quantum, do you recall which noncombat you got on 3 turns remaining?

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

Post by QuantumNightmare » Mon Sep 03, 2007 5:28 pm

Never mind. I checked my mafia logs, and I got the noncombats at turns 7,4,1. Nothing to see folks!

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

Post by QuantumNightmare » Mon Sep 03, 2007 6:46 pm

Double post. I've updated this sheet to be able to graph stat gains across changing conditions. So far this is a developers-only feature, but here are some interesting conclusions so far.

I've graphed stat gains for a moxie class with +20% noncombats, +30% item drops and a 30 pound volleyball, as +ML varies from 0 to +125ML. The sheet calculates the anti-synergy between ML and combat modifiers, and gives some interesting results. Here are graphs done at 70 mainstat, 104 mainstat, and 125 mainstat (only moxie gains are graphed).

70 Mainstat
Image
104 Mainstat
Image
125 Mainstat
Image

In each case, the sign gym gives very low gains and absinthe gives very high gains. What is important is, +ML only hurts gallery, palindome, castle, and high level ballroom. In all other locations, it's more worth it to just pump +ML and ignore combat modifiers! (Note that at +100ML, all combat modifiers are canceled out and so all areas start to improve)

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

Post by Flolle » Mon Sep 03, 2007 8:05 pm

QuantumNightmare wrote:What is important is, +ML only hurts gallery, palindome, castle, and high level ballroom. In all other locations, it's more worth it to just pump +ML and ignore combat modifiers!

So I was right? Cool. Yay for confirmation from the different approach! :D (though as noted before, our Ballrooms are different => mine isn't going into the negative)

And as far as the high level Ballroom goes, that depends on how much +itemdrop your running, or am I thinking wrong here?

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

Post by QuantumNightmare » Mon Sep 03, 2007 8:46 pm

Flolle wrote:And as far as the high level Ballroom goes, that depends on how much +itemdrop your running, or am I thinking wrong here?
It's a combination of many things.

High mainstat -> higher noncombat yields.
High Item drop -> higher combat yields (which in turn depend on mainstat...)

Increased +ML -> higher combat yields, fewer noncombats. So yes, the more +items you are running, the more likely you are to prefer adding +ML than adding +noncombats.

DarthDud
Inscrutable Pi
Posts: 265
Joined: Fri Apr 06, 2007 12:41 am

Post by DarthDud » Tue Sep 04, 2007 1:44 pm

This would only be particularly useful if you limited the stat gains to moxie substats, ignoring the other two which aren't very important...

Also, how are you calculating absinthe gains? Based on average # turns in the wormwood?

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

Post by QuantumNightmare » Tue Sep 04, 2007 11:32 pm

I'm already only graphing moxie gains in those 3 graphs. Total stats really....don't matter at all.

And as for absinthe stats, I explained in an earlier post. Considering you can get noncombats on turns 9-7, 5-4, and 1... you'll spend between 3 and 6 adventures in the wormwood. That's an average of 4.5, with 3 noncombats giving 1*mainstat and 1.5 combats at ML = moxie (+ML has no effect here). The gains from those 4.5 turns are averaged per turn, and the other 5.5 turns are expected to be spent in a better leveling location.

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

Post by QuantumNightmare » Sat Sep 08, 2007 10:39 am

Big changes.

- Sheet will now calculate every zone both with and without +ML applied (MCD, if used, is always applied). It will then display whether you get more of the desired stat with or without +ML, and will display stat gains for the best case scenario ONLY.
- This means if you add a +ML of 90, only some of the zones will display stat gains with +ML, the rest will display using only the MCD and applied noncombat boosts.
- Macros are finally all gone.

This update took a lot of behind the scenes changes, so I might have broken something. But it looks good to me!

dalgar
Pie of the Sky
Posts: 26
Joined: Sun Sep 02, 2007 9:32 am

Post by dalgar » Sat Sep 08, 2007 2:48 pm

Hey QN,

What does antipode mean, and what does it do?

And many thanks for the great spreadsheet!

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

Post by QuantumNightmare » Sat Sep 08, 2007 3:00 pm

Well, the main job of antipode is to completely break the spreadsheet. Oops! New version is released as of 2 minutes ago.

Otherwise antipode = antiphon the AT buff, misspelled pretty badly. +1 stats per combat per statistic. It's one of the many yes/no choices on the sheet.

Also.... ballroom now attempts to pickpocket dance cards as a moxie class. This assumes you get initiative on every waltzer. As well, Harold's Bell option now works for the ballroom. The combination of harold's bell and an odor extractor is....powerful.

charred_hcn

Post by charred_hcn » Sun Sep 09, 2007 1:50 pm

just found this from the kol forums, posted this there and then found out i can post here without registering.


just downloaded this, and it is borked for sombreros. on the calculation pages the battlefield calcs are trying to calculate the mainstats types. works for other fams tho

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

Post by QuantumNightmare » Sun Sep 09, 2007 3:15 pm

Sombrero is fixed. Looks like I broke it when I added some new features over the past few days. Thanks for pointing that out!

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

Post by Flolle » Sat Sep 15, 2007 9:57 pm

Sorry if I missuse this thread again, but I finally found out why our Ballrooms are different.

Whenever a quartet is rolled and cannot occur due to its 20 turns counter I treated it like a skipped noncombat, which means that whether the current encounter is a combat or noncombat is rolled again. Just now I tested it with rerolling a noncombat and I got almost the same values as you did with your math excursion on the first page. Happiness! Very. Much. Happiness! :D

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

Post by QuantumNightmare » Sat Sep 15, 2007 11:10 pm

Great! Wasn't the castle/bathroom/ballroom the only problems we had? Bathroom was my mistake, and you've found the problem in your ballroom. If the same mistake was made while skipping the wheel... we might finally have consistent results!

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

Post by Flolle » Sun Sep 16, 2007 12:48 am

QuantumNightmare wrote:Great! Wasn't the castle/bathroom/ballroom the only problems we had? Bathroom was my mistake, and you've found the problem in your ballroom. If the same mistake was made while skipping the wheel... we might finally have consistent results!

Castle mainstat and Palindome moxie are still off by a significant margin. Bathroom differences are now relatively small. But no idea why that is the case, because all the code used for those seems to work elsewhere. E.g. if noncombat skipping didn't work the Ballroom would still be different and if noncombat stats didn't work, Gallery, Ballroom and Bedroom wouldn't fit almost perfectly.

I had so many problems with the Ballroom mainly because it had much relatively complicated stuff that simply only appeared there: Dance Cards and Quartet mechanic. But hey, on the plus side, if Jick ever decides to implement some other places with these mechanics, the code will be there for it.

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

Post by Flolle » Sun Sep 16, 2007 1:10 pm

I believe I found out what's wrong with your Palindome.

You're capping the moxie noncombat wrong (field K12 and K40 in the calculations sheet). It says the cap is at 250, but 100 would actually be the right value. So I changed it to 100 and turns out our outputs are the same now.

(I haven't pushed out the ML changes in the Palindome yet, so don't bother testing it with the Simulator version on my website, I have a newer one on my PC. ;)).

I haven't looked at the Castle yet, but that one is probably going to be more complicated. Ah well, first to check what my adventure distribution looks like and then ... *walks away mumbling*

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

Post by QuantumNightmare » Sun Sep 16, 2007 1:37 pm

Really? I'm pretty sure the moxie noncombat caps at 250, while the pep talk caps at 100. And the pep talk is only optimal between 0-200 mainstat, after which point it's best to take papayas and trade them in. This trading mechanism caps at 300 mainstat.

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

Post by Flolle » Sun Sep 16, 2007 1:54 pm

That's strange. The wiki says what you say. So I just went to the palindome to check and got 222 moxie substats with mainstat 221. This really is ... where the hell did I get the 100 from!? o.O

Anyway, problem solved, my cap was wrong. Either way, our tools have the same output for the palindome now.

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

Post by QuantumNightmare » Sun Sep 16, 2007 4:08 pm

Alright, so now we only have the castle as a difference. We have the exact same offstats, with a difference of 4-5 in the mainstat. Looks like the guilty party is the noncombat distribution, or noncombat stat gain. My equation for the castle noncombat gains is:

2.25*mainstat capped at 250.

With the wheel set properly, there are only 3 noncombats that will take up an adventure. Out of those 3, the stat noncombat comes twice as often. In other words, half of noncombats that are not skipped will be stat adventures, the other half will be wasted.

Also: new version. I've updated the palindome ML, fixed the castle and gallery meat gains from noncombats, and absinthe is no longer highlighted even if it is the best stat location.

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

Post by Flolle » Sun Sep 16, 2007 4:23 pm

DANGER! LONG COMPLICATED POST AHEAD! YOU HAVE BEEN WARNED!


As I said before, I went ahead and looked at the encounter rate in the Castle from the Area Simulator. I noticed that the ratios between the noncombats where wrong, so I checked the source and couldn't find a fault in it. But then I had a hunch and turned off the queue and the consequence of it was getting the same numbers as you do.

I must say that this bothers me.

I'm pretty sure I didn't do anything wrong while implementing the queue, so here is a question to all mathy people: Does the queue work as an equalizer in locations where encounters of the same kind (e.g. only combats or noncombats) have different appearance rates?


To give an example to show what I mean:

Say we have a location with 4 combats.

Combat A has an appearance rate of 70%
Combat B has an appearance rate of 10%
Combat C has an appearance rate of 10%
Combat D has an appearance rate of 10%

So the ratio between the combats should be 7:1:1:1.

But when you're actually adventuring in the location this means that combat A gets put into the queue very often and probably is in it most of the time. All other combats are not that often in the queue. So combat A gets rerolled very often. Most of the time it will not change the overall ratio between the combats, but whenever one of the combats is not in the queue, it is much more likely to get through the encounter rolling process and thus changing the overall ratio.

In the end this wouldn't change the ratio between combat B, C and D, but combat A wouldn't have a ratio of 7 anymore, because as I said above, it is in the queue most of the time and thus cannot profit from the above described encounter rerolling. So in the end the overall ratio would be along the lines of (something below 7):1:1:1. And the average appearance rates would be something like this:

Combat A (something below 70)%
Combat B (something above 10)%
Combat C (something above 10)%
Combat D (something above 10)%

So the conclusion is that combat A appears less often than what would be expected while all other combats appear more often. And the consequence from that would be that quite a few appearance rates spaded are not the actually hardcoded appearance rates of the encounters. They are appearance rates which already take the queue into account. That could potentially screw up turn calculations for things like queueing up the GmoB.

And this would explain why our Castles and Bathrooms are still different, QN. I'm basically doubling the effect of the queue on these locations and am thus screwing up appearance rates.

Ok, now that this far too complicated post is over, I'm going to hope that at least one person with the necessary math skills understood what I meant and can tell me whether everything above is valid or simply both, a very very big brainfart and a fault in the source of the Area Simulator. ;)


Edit: I looked at the wiki page and have a feeling that at least part of the problem is addressed here. But I must admit that that is a bit too complicated for me(one of the reasons I programed a Simulator, to sidestep these math problems). Again, mathy people please step up. ;)

Edit edit: And from the looks of this, I am in fact right. Other people noticed one year before me(why didn't anybody tell me about it? T.T). But I haven't seen a formula to find the raw appearance rates. Uhh. :?

User avatar
Mad Hamish
AFH
Posts: 1517
Joined: Mon Apr 09, 2007 8:03 pm

Post by Mad Hamish » Thu Sep 20, 2007 12:37 pm

Does your spreadsheet take account of the meat yield from the wallet and coins in the bedroom when you select to buy and use clovers?
Image
Hi-keebated again? No way!

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

Post by QuantumNightmare » Thu Sep 20, 2007 7:04 pm

So far, I only calculate meat from the gallery, castle, and palindome. So those are the only zones that use the "buy clovers" option. I'll add the rest eventually, and adding support for meat from the bedroom (wallets and everything else) is on my spading list.

User avatar
Mad Hamish
AFH
Posts: 1517
Joined: Mon Apr 09, 2007 8:03 pm

Post by Mad Hamish » Thu Sep 20, 2007 7:25 pm

That's kinda what I figured. So I decided to stick with the absinthing/bedroom combo through today, despite that the sign gym is marginally better than the average between them.
I really need to write an absinthing script before I do this again.
Image
Hi-keebated again? No way!

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

Post by QuantumNightmare » Sat Oct 13, 2007 12:28 am

New update. Let's call this version a beta, as there are probably a few bugs to be worked out with the meat calculations.

-Middle chamber is now calculated (due to it's meat drops)
-Bedroom meat drops are calculated.
-Bedroom assumes that if not myst class, you'll go for the 500 meat instead of the offstats
-NPZR farming makes a small appearance. Calculations are made on a maximum 29 round combat encounter.

The point of calculating the NPZR is to help quantify meat farming, with the express purpose of using this meat to buy clovers and power-level.

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

Post by QuantumNightmare » Sun Nov 18, 2007 7:23 pm

Minor changes. The zones are now grouped more intuitively, and the ed calculations are finally displayed. Because people still seem to think he's good for stats.

Post Reply