Box of Fireworks

Public forum for the discussion of strategy.
Post Reply
Kevlar
Inscrutable Pi
Posts: 334
Joined: Fri Apr 06, 2007 4:04 am

Box of Fireworks

Post by Kevlar » Thu Jul 03, 2008 8:46 am

3 days into the month and I'm surprised not to have heard them mentioned on here. I guess Hobopolis totally eclipsed them.

What are your early impressions like? I'm in no real hurry to rush out and buy one yet as my pumpkin bucket is getting me through the hobopolis.

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

Post by Alrik » Thu Jul 03, 2008 9:06 am

Looks like another mayflower to me. Since it has a method for leveling your familiar more quickly (kills by roman candle) and can provide bonus stats at the end of the fight, I'd like to see a head-to-head between the two.

Though I haven't played with the fireworks yet, it feels like the mayflower is less consistent -- it provides more different types of bonuses but will be less reliable about providing the ones you want -- so the fireworks -may- be better for speed ascenders.

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

Post by QuantumNightmare » Thu Jul 03, 2008 9:40 am

GoldS has some nice data on the wiki. This thing *always* fires, for either +3 stats or +20 to 25% items. So it's a pumpkin bucket with half a wax lip attached. Looks like a nice compromise between the two items. +familiar xp looks useless.

Wax lips are better for pure stats but the fireworks give close to the same bonus along with a significant amount of +items, especially used on a pixie or other combo familiar. I like it.

LegendaryFerret
AFH
Posts: 164
Joined: Tue Jan 15, 2008 8:25 pm

Post by LegendaryFerret » Thu Jul 03, 2008 10:26 am

I'm with QN that familiar experience looks worthless, since getting it conflicts with getting the bonus stats from a V mask.

Other than that, it's pretty close to wax lips (on a stat familiar going from 15 to 20 pounds is .5 stats, so it's 2 stats rather than 3), with the nice perk of a large amount of item drops and buffing your item drop familiars. On the one hand, 1 stat a combat over the course of an entire ascension adds up (say 500 substats). On the other, that's only 1.5 pulls, so if the extra item drop saves you 2 pulls, you're ahead of wax lips. I'm not certain yet which I like better, but I think it'll be close.

Compared to the PPB, I don't think there's really a contest: this wins hands down. For almost the same item drop, I can get 1.5 stats per turn and have control of the damage at the start of the fight? Sure, in a couple of situations (hobo boss, say), I'd rather have a PPB (or even more a maypole, but I don't own one of those), but in general, I think the PPB's pretty much outclassed.

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

Post by KujjieKujjieKoo » Thu Jul 03, 2008 10:31 am

Wait... it's a half PPB, half wax lips, *and* 5-lbs? That seems... really good.
Image

LegendaryFerret
AFH
Posts: 164
Joined: Tue Jan 15, 2008 8:25 pm

Post by LegendaryFerret » Thu Jul 03, 2008 10:34 am

And familiar leveling if you kill stuff with its skill.

And it lets you get the jump sometimes when you wouldn't otherwise.

(and it's half-maypole, so half-twice PPB)

salien
Pie in the Sky
Posts: 115
Joined: Fri Jan 11, 2008 1:50 pm

Post by salien » Tue Jul 08, 2008 7:12 pm

Yeah, I think Hobopolis has distracted just about everyone from how good this is. On a sombrero (now that the formula is solved, yay!), the 5 lbs is often a full stat point at decent ML (I say often because of the flooring), making it 2.5 stats vs. 3 stats from lips. That's only maybe 200 stats on a speed run, only 100 mainstat; if you're moxie and leveling in the ballroom, the +item and extra shaman/pixie pounds will easily make that up. Then the extra +items everywhere else, and extra pounds for things like hound dog and meat fam are just bonus. Even if your main fam is a volley, that's still only 200 mainstat, very likely made up with extra dance cards. If you're not moxie, the stats are a loss, but can probably be made up in 2 or so turns, vs. the turns saved by the extra item drops, extra hound dog weight, extra meat weight, generally increased flexibility from having +5 lbs on anything you use...

As for the bucket, this thing is better in pretty much every possible circumstance, which is kind of absurd.

The main point in favor of the lips is that most of the lost stats will happen on day 1, when they matter most; still, I highly doubt the stats will even make a full turn's difference.

Also, for the +fam xp, it could be worth messing with after hitting level 12; an extra pound or two on a hound dog or meat fam at that point could matter more than +4 stats from the mask. (At the least it's less of a pain in the ass.)

At the very least, absolutely anyone interested in speed ascension who doesn't have wax lips needs the fireworks.

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

Post by Seal Lubber » Wed Jul 09, 2008 3:12 am

I fought Oscus with +378% item drops and a little box of fireworks, getting 4 jars of pickle juice. If the 1 item/100% drops is true, the box gave 22% or better.

slackerjack
Just Plain Old Pie
Posts: 14
Joined: Wed May 21, 2008 10:27 pm

Post by slackerjack » Wed Jul 09, 2008 12:28 pm

salien wrote:At the very least, absolutely anyone interested in speed ascension who doesn't have wax lips needs the fireworks.


Indeed.

I'm not yet convinced that BoF is strictly better, but that doesn't really matter at this point. Practically speaking, BoF seems good enough to serve as a full substitute for lips -- not just another workable-but-clearly-weaker alternative like the PPB and Bouquet.

I'll probably get one myself, since I wouldn't be surprised if it manages to edge out the lips after more detailed analysis/experience.

User avatar
lostcalpolydude
Forum Maestro
Posts: 2940
Joined: Wed Apr 23, 2008 6:54 pm

Post by lostcalpolydude » Sat Jul 19, 2008 3:10 pm

In the thread on the main forum about this item, there was some discussion about how useful the +item drop is compared to other +item stuff. Specifically, there was discussion about whether having +25% drops half the time is the same as having +12.5% drops all the time, ignoring the case where the +25% drops causes an item's drop rate to go above 100%. I was making the case that the variance was higher for the little box of fireworks compared to something like the PPB (ignoring the fact that the PPB is slightly more than half the fireworks value), making the comparison to the greater variance of !pipes over feathers. Another person was saying that the variance from both types of +item effects is identical. I made this post, and got a PM response:
BigBadBongo wrote:I'd rather not spam the thread with more corrections, would be nice if you could edit your post so it was no longer misleading to people trying to work out best thing to do. I'm not asking you to say you're wrong, just to remove any misleading information and maybe put in the correct information if you feel you understand. I don't think anyone quoted you.

"Except you're getting your terminology mixed up. The average value is the same, the variance is not the same. There's more chance for the RNG to decide to be nice and give you the drop sooner, and there's also more room for the RNG to decide it hates you and make the drop take longer. It averages out, but the difference between best-case and worst-case scenarios is larger."

No, the variance is the same, I am quite familiar with the difference between the terms!

First Y items in X adventures. P1/V1 assume 50% chance of 25%, P2/V2 are a constant 12.5%. D is base item drop rate, E is item drop modifier before these bits of familiar equipment.

A way to calculate variance:
V = [ Mean of Square ] - [ Square of Mean ]

[ Mean of Square ] = [ Sum{n=1 to X}( P(n drops)*(n^2) ) ]
[ Square of Mean ] = [ Sum{n=1 to X}( P(n drops)*n ) ]^2

Agreed?

So as long as P1(n drops) = P2(n drops) for all n then V1=V2. Agreed?

P2(n drops) = [X!/n!(X-n)!]*(P2(1 drop in 1 adventure)^n)*(P2(no drop in 1 adventure)^(X-n)) = [X!/n!(X-n)!]*((D*(1+E+0.125))^n) * ((1-D*(1+E+0.125))^(X-n))


P1(n drops) = [X!/n!(X-n)!]*(P1(1 drop in 1 adventure)^n)*(P1(no drop in 1 adventure)^(X-n))
= [X!/n!(X-n)!]*((0.5*[drop with bonus] + 0.5*[drop without bonus])^n)*((0.5*[no drop with bonus] + 0.5*[no drop without bonus])^(X-n)) //Substitute in actual probability functions//
= [X!/n!(X-n)!]*((0.5*[D*(1+E+0.25)] + 0.5*[D*(1+E)])^n)*((0.5*[1 - D*(1+E+0.25)] + 0.5*[1 - D*(1+E)])^(X-n)) //Substitute in actual probabilities//
= [X!/n!(X-n)!]*((0.5*D*1+0.5*D*E+0.5*D*0.25 + 0.5*D*1+0.5*D*E)^n)*((0.5*1 - 0.5*D*1 - 0.5*D*E - 0.5*D*0.25 + 0.5*1 - 0.5*D*1 - 0.5*D*E))^(X-n)) //Multiply out brackets//
= [X!/n!(X-n)!]*((D+D*0.125+D*E)^n)*((1 - D - D*E - D*0.125 ))^(X-n)) //Collect like terms and multiply multiple coefficients//
= [X!/n!(X-n)!]*((D*(1+E+0.125))^n) * ((1-D*(1+E+0.125))^(X-n)) //Take a factor of D outside some brackets//
= P2(n drops) //Refer to calculation of P2 above//

So for all n P1(n drops) = P2(n drops) therefore V1(number of drops)=V2(number of drops).

So variance of # of drops in X adventures is identical for the two mechanics.

Now for variance in # of adventures before the first drop.

V = [ Mean of Square ] - [ Square of Mean ]

[ Mean of Square ] = [ Sum{n=1 to infinity}( P(n adventures)*(n^2) ) ]
[ Square of Mean ] = [ Sum{n=1 to infinity}( P(n adventures)*n ) ]^2

So it suffices to show that P1(n adventures) = P2(n adventures).

Agreed?

P2(n adventures) = P2(no drops in n-1 adventures)*P2(drops on nth adventure)
= [(P2(doesn't drop))^(n-1)]*[P2(drops)]
= [(1-D*(1+E+0.125))^(n-1)]*[D*(1+E+0.125)]

P1(n adventures) = P1(no drops in n-1 adventures)*P1(drops on nth adventure)
= [(P1(doesn't drop))^(n-1)]*[P1(drops)]
= [(0.5*(1-D*(1+E)) + 0.5*(1-D*(1+E+0.25)))^(n-1)]*[0.5*(D*(1+E)) + 0.5*(D*(1+E+0.25))]
= [((0.5-0.5*D-0.5*D*E) + (0.5-0.5*D-0.5*D*E-0.5*D*0.25))^(n-1)] * [0.5*D+0.5*D*E + 0.5*D+0.5*D*E+0.5*D*0.25]
= [(1-D-D*E-D*0.125)^(n-1)] * [D+D*E+D*0.125]
= [(1-D*(1+E+0.125))^(n-1)]*[D*(1+E+0.125)]
= P2(n adventures)

So the variance in the number of adventures required to get the first one is the same.

All means and variances of individual items are the same. All probabilities of getting a certain number of items in a certain number of adventures are the same.

Essentially if you remove the message that tells you whether or not you get the item drop and examine only one item (the item you are looking for) then you will be totally and utterly unable to tell the difference between the 2 items. *Everything* about them behaves the same.

Disclaimer: This is, as stated previously, only valid for total item drop chances of no more than 100%, which is almost always the case. Only very very rarely will this break down.

At this point, some of this stuff goes over my head because I haven't had enough experience with probability to work through it (I can't even tell if the initial premise is correct, so I haven't bothered wading through the later math). I figured someone here could go through and see if it is correct or find why it isn't correct. I'll point that person over to this post also.
bmaher: Softcore character
lostcalpolydude: Hardcore character (abandoned)

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

Post by Flolle » Sat Jul 19, 2008 3:48 pm

Well, no math from my side, because that post also goes over the top of my head (ok, truth is, I didn't even try to understand because I'm feeling lazy and still need to play my turns ;)), but I checked this with the Area Simulator and results for both mean and standard deviation were the same with both setups.

User avatar
deusnoctum
Statistically likely to return
Posts: 820
Joined: Wed May 07, 2008 4:00 pm

Post by deusnoctum » Sat Jul 19, 2008 4:06 pm

Flolle wrote:Well, no math from my side, because that post also goes over the top of my head (ok, truth is, I didn't even try to understand because I'm feeling lazy and still need to play my turns ;)), but I checked this with the Area Simulator and results for both mean and standard deviation were the same with both setups.


I'm having a hard time following the math, particularly in a crappy text format, but since stdev = sqrt(variance), if the stdevs are equal, the variances are equal. Intuitively, though, it feels like he's right.

User avatar
Pet Rock Steve
Pie in the Sky
Posts: 111
Joined: Thu Apr 03, 2008 5:00 pm

Post by Pet Rock Steve » Sun Jul 20, 2008 12:35 am

The math looks good to me, and a couple Monte Carlo simulations agree as well.

Oxbarn
Just Plain Old Pie
Posts: 18
Joined: Tue May 06, 2008 9:55 pm

Post by Oxbarn » Sun Jul 20, 2008 11:32 am

When I think about it, I have a hard time seeing how the results could be not the same. There's really only one number involved, which is the final chance to get the item from an adventure. It just can't matter how that percentage was arrived at, whether you call the die "item drop" or "firework roll followed by item drop" or "getting the right adventure followed by item drop".

It's like, there can't be any difference between getting a 100% drop from an adventure that occurs 1/5 of the time, and getting an item that drops at 20% from every adventure. It means nothing that the "item drop" percentage is constant (20%, PPB-style) vs. spikey (100% or zero, fireworks-style). You've still got to roll the exact same one-in-five chance in either scenario.

BigBadBongo
Just Plain Old Pie
Posts: 1
Joined: Sun Jul 20, 2008 1:30 pm

Post by BigBadBongo » Sun Jul 20, 2008 1:35 pm

Oxbarn wrote:It's like, there can't be any difference between getting a 100% drop from an adventure that occurs 1/5 of the time, and getting an item that drops at 20% from every adventure. It means nothing that the "item drop" percentage is constant (20%, PPB-style) vs. spikey (100% or zero, fireworks-style). You've still got to roll the exact same one-in-five chance in either scenario.


This is exactly correct, although many people seem unable to grasp that probability rolled in two different stages is the same as probability rolled all at once.

My principle worry was that too many people would see the multiple misleading posts in the KoL forum thread and be led to take actions they would otherwise not do, so I've been working to try to reduce that. Hence my message.

Incidentally I apologise for the mess of text, but I'm no wizard with the internets.

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

Post by DarthDud » Thu Jul 31, 2008 12:20 pm

Oxbarn wrote:When I think about it, I have a hard time seeing how the results could be not the same. There's really only one number involved, which is the final chance to get the item from an adventure. It just can't matter how that percentage was arrived at, whether you call the die "item drop" or "firework roll followed by item drop" or "getting the right adventure followed by item drop".

It's like, there can't be any difference between getting a 100% drop from an adventure that occurs 1/5 of the time, and getting an item that drops at 20% from every adventure. It means nothing that the "item drop" percentage is constant (20%, PPB-style) vs. spikey (100% or zero, fireworks-style). You've still got to roll the exact same one-in-five chance in either scenario.

It can matter, but not often. Unless I am thinking wrong with the following thought experiment.

Let's say you're doing the goatlet, and you have, oh, +137.5% item drops.

With a constant +12.5%, you will always get a goat cheese. With a variable +0/25%, 3 out of 6 goats you encounter will have a 5% chance of not dropping cheese! This increases your expected time to completion.

Anyways, outside of such situation (which almost never happen), I agree that there is no difference.

Oxbarn
Just Plain Old Pie
Posts: 18
Joined: Tue May 06, 2008 9:55 pm

Post by Oxbarn » Thu Jul 31, 2008 10:27 pm

lostcalpolydude wrote:ignoring the case where the +25% drops causes an item's drop rate to go above 100%.

probably got lost in the flood of math, but yes.

User avatar
Fred Nefler
AFH
Posts: 812
Joined: Sat Jan 17, 2009 5:47 pm

Post by Fred Nefler » Sat Feb 21, 2009 1:35 am

Pardon me while I necro this thread to provide a more elegant and formal proof that +25% half of the time is equivalent to +12.5% all of the time. My professional training commands it.

Math powers activate!

NB: The 25%/12.5% pair is no special thing. +p% half the time and +p/2% all the time are the same for all p. Excluding the previously mentioned pathological cases of guaranteed drops (or gauranteed no drops). Nor is the half thing, for that matter. 50% 1/4 of the time is the same as 12.5% all the time, etc., provided the same pathologies do not arise.

P.S. Yes, I know there's a period missing in that file. Deal with it!

User avatar
Eleron
Has a thing!
Posts: 870
Joined: Tue Jun 03, 2008 1:52 pm

Post by Eleron » Sun Feb 22, 2009 8:09 am

Fred Nefler wrote:A more elegant and formal proof that +25% half of the time is equivalent to +12.5% all of the time.
To reach that conclusion from your proof, you need to assume that the utility of all the items dropped in a turn are independent, which isn't always true.

Consider the super-blooper, which drops N white pixels, each with a 10% chance, and you'd like to collect N. You also have the lucky box that has a 1% chance of giving +900% item drops, compared to the mehpole that gives a flat +9% item drops.

For N = 1, the chance of getting N pixels on the first turn is 10.9% for the box and 10.9% for the mehpole.
For N = 2, the chances are 1.99% for the box and 1.1881% for the mehpole.
For N = 1000, the chances are ~1% for the box and ~0% for the mehpole.

The expected number of drops per turn is naturally the same for each alternative, but the probability distribution is skewed. This also affects the probability distribution and expected value(!) for number of turns to collect M items.

*hug*

Oxbarn
Just Plain Old Pie
Posts: 18
Joined: Tue May 06, 2008 9:55 pm

Post by Oxbarn » Sun Feb 22, 2009 7:01 pm

OK, that's a pretty funny and clever result. Not actually relevant to the current game, since only farmers ever face this choice (fireworks being obviously better for ascension) and for farmers drop utility is as nearly independent as one could ask for, but cute!

I guess a more realistic example is that if you're farming the last piece of an outfit and you have a wand, then the pumpkin bucket is just slightly better, because some of the fireworks' spiky bonus could be "wasted" on giving you multiple drops when you only needed one. (The bucket is more likely to spread these multiple drops over multiple turns, letting you stop after the first instead of waiting for a spike.)

User avatar
Fred Nefler
AFH
Posts: 812
Joined: Sat Jan 17, 2009 5:47 pm

Post by Fred Nefler » Sun Feb 22, 2009 7:41 pm

Eleron wrote:To reach that conclusion from your proof, you need to assume that the utility of all the items dropped in a turn are independent, which isn't always true.

I'm not sure how you would even begin to measure "utility" in a proper fashion.

The actual issue, as I would put it, is that the sum of multiple random variables that depend on the selector random variable S is no longer identically distributed to a sum of random variables that are identically distributed but not dependent on the same selector. This is because the same value of S is used for each term in the sum, rather then a new independent copy of S. That's for drops in the same turn. For a sum of variables in distinct turns that each depend on S, the sum is identically distributed to those not dependent on a selector, since the selector is independently determined for each turn.

An issue I kind of intentionally didn't consider, because its relevance is situation-specific. But I have to conceed that I might as well have gone ahead and mentioned it.

User avatar
Eleron
Has a thing!
Posts: 870
Joined: Tue Jun 03, 2008 1:52 pm

Post by Eleron » Mon Feb 23, 2009 9:48 am

Fred Nefler wrote:I'm not sure how you would even begin to measure "utility" in a proper fashion.

It's not necessary to assign actual values as long as you can determine whether they're independent. Multiple equivalent drops from one monster usually aren't.

U(x) + U(y) = U(x + y) normally isn't strictly true even for x != y, but the big violation is U(x) + U(x) = U(x + x) which is often blatantly false.

Fred Nefler wrote:An issue I kind of intentionally didn't consider, because its relevance is situation-specific. But I have to conceed that I might as well have gone ahead and mentioned it.

So what you're saying is that given some semi-reasonable assumptions, then the hypothetical item bonuses are equivalent for monsters with only one possible drop. That's true, but a significantly weaker statement =)

User avatar
Bantler
Just Plain Old Pie
Posts: 18
Joined: Fri Apr 24, 2009 10:25 am

Post by Bantler » Tue May 05, 2009 12:15 pm

The frumious bandersnatch makes fireworks obsolete?
Calms the moose. You should try it sometime.

User avatar
lostcalpolydude
Forum Maestro
Posts: 2940
Joined: Wed Apr 23, 2008 6:54 pm

Post by lostcalpolydude » Tue May 05, 2009 12:58 pm

Bantler wrote:The frumious bandersnatch makes fireworks obsolete?
Not obsolete, but less useful. You're using the bander a lot of the time, and you use its equipment with it. Fireworks are still worth a pull due to the extra +items, +meat, and +MP regen you get on the associated familiars, but you don't lose quite as much from not having fireworks. They're still cheap enough that it doesn't make much sense not to have and use them though.

On a 2 day SCNP run, fireworks are probably obsolete, though I don't have the experience to say. (Ignoring the crazy run by Pantsless where he did weird things out of necessity.)
bmaher: Softcore character
lostcalpolydude: Hardcore character (abandoned)

User avatar
Bantler
Just Plain Old Pie
Posts: 18
Joined: Fri Apr 24, 2009 10:25 am

Post by Bantler » Tue May 05, 2009 1:38 pm

Doesn't the mayflower bouquet edge it out with drops that can convert pounds to bonus frumious bandersnatch runaways? Who is going to keep updated on this stuff if quantum nightmare is gone?
Calms the moose. You should try it sometime.

User avatar
lostcalpolydude
Forum Maestro
Posts: 2940
Joined: Wed Apr 23, 2008 6:54 pm

Post by lostcalpolydude » Tue May 05, 2009 9:29 pm

Doubling the mainstat xp by wearing the bander's equipment is much better than trying to get the bander to be quadruple its buffed weight by using a bouquet. The extra stats from using the bander's equipment should easily beat the extra runaway that you might get from using a bouquet (or even the extra runaway you might get from fireworks kills).
bmaher: Softcore character
lostcalpolydude: Hardcore character (abandoned)

rottingflesh
Inscrutable Pi
Posts: 300
Joined: Sun May 04, 2008 12:30 am

Post by rottingflesh » Tue May 05, 2009 11:58 pm

I think he means equipping the bouquet on everything but the bander.

User avatar
Bantler
Just Plain Old Pie
Posts: 18
Joined: Fri Apr 24, 2009 10:25 am

Post by Bantler » Wed May 06, 2009 10:02 am

I also mean using the bouquets flower drop with the familiar weight in conjunction with the frumious bandersnatch for an extra runaway a day. Do you lads have a secret forum or something where you hide the good stuff?
Calms the moose. You should try it sometime.

User avatar
stupac2
Oh my! Guy with Pie!
Posts: 3027
Joined: Mon Dec 08, 2008 10:04 pm
Location: Stanford, CA
Contact:

Post by stupac2 » Wed May 06, 2009 10:15 am

Bantler wrote:I also mean using the bouquets flower drop with the familiar weight in conjunction with the frumious bandersnatch for an extra runaway a day. Do you lads have a secret forum or something where you hide the good stuff?
Yes, but in this case no. I'd rather have the 12.5% items and 1.5 stats than a shot at a free runaway.

User avatar
Bantler
Just Plain Old Pie
Posts: 18
Joined: Fri Apr 24, 2009 10:25 am

Post by Bantler » Wed May 06, 2009 10:24 am

The frumious bandersnatch is hogging all the stats anyway and toggle 12% is pretty much shit isn't it? Bouquet gives some bonus pounds on other familiars and drop rate flowers when you need them. Doesn't a runaway every day add up to better stats and item chances than you get from fireworks?
Calms the moose. You should try it sometime.

User avatar
lostcalpolydude
Forum Maestro
Posts: 2940
Joined: Wed Apr 23, 2008 6:54 pm

Post by lostcalpolydude » Wed May 06, 2009 10:47 am

Bantler wrote:toggle 12% is pretty much shit isn't it?
Getting +25% items half the time is "basically" the same as getting +12.5% items all the time. See the math and discussion above for why that is, or be like me and trust that all the people claiming to understand the math really do understand it. Also, just because a lot of your stats come from the bander doesn't mean you don't want stats when you aren't using it.
bmaher: Softcore character
lostcalpolydude: Hardcore character (abandoned)

User avatar
Bantler
Just Plain Old Pie
Posts: 18
Joined: Fri Apr 24, 2009 10:25 am

Post by Bantler » Wed May 06, 2009 11:29 am

In a fast 500 turn run there are about 100 non frumious bandersnatch combat turns. So that's 75 main stat points over an entire run with the fireworks. How does that beat 2 or 3 runaways?
Calms the moose. You should try it sometime.

User avatar
stupac2
Oh my! Guy with Pie!
Posts: 3027
Joined: Mon Dec 08, 2008 10:04 pm
Location: Stanford, CA
Contact:

Post by stupac2 » Wed May 06, 2009 1:30 pm

Bantler wrote:In a fast 500 turn run there are about 100 non frumious bandersnatch combat turns. So that's 75 main stat points over an entire run with the fireworks. How does that beat 2 or 3 runaways?
And 12.5% items. That beats some percent chance at a a 60% chance of an extra free runaway.

rottingflesh
Inscrutable Pi
Posts: 300
Joined: Sun May 04, 2008 12:30 am

Post by rottingflesh » Wed May 06, 2009 2:23 pm

The Bouquet has a 12.5% chance of each of these effects firing at the end of a combat:
1. +3 Familiar Exp
2. +40% meat drops
3. +5 substats!!!!
4. some amount of +items, currently unspaded, but I'd guess 20%.

You also get some number of flowers each day. The only truly relevant ones granting +3 familiar weight and +30% meat drops.

It also helps you v-crisis at higher ML with it's navel-like blocks.
I think he may be onto something.


...Maybe.
Last edited by rottingflesh on Wed May 06, 2009 5:23 pm, edited 1 time in total.

User avatar
kindia
Pie in the Sky
Posts: 156
Joined: Thu Aug 28, 2008 2:34 pm

Post by kindia » Wed May 06, 2009 4:46 pm

rottingflesh wrote:The Bouquet has a 12.5% chance of each of these effects firing at the end of a combat:
1. +3 Familiar Exp
2. +40% meat drops
3. +5 substats!!!!
4. some amount of +items, currently unspaded, but I'd guess 20%.

You also get some number of flowers each day. The only truly relevant ones granting +3 familiar weight and +30% meat drops.

I think he may be onto something.


...Maybe.
so does that mean the mayflower doesnt fire every time?
at least the LBoF fires every time.

User avatar
stupac2
Oh my! Guy with Pie!
Posts: 3027
Joined: Mon Dec 08, 2008 10:04 pm
Location: Stanford, CA
Contact:

Post by stupac2 » Wed May 06, 2009 7:42 pm

rottingflesh wrote:The Bouquet has a 12.5% chance of each of these effects firing at the end of a combat:
1. +3 Familiar Exp
2. +40% meat drops
3. +5 substats!!!!
4. some amount of +items, currently unspaded, but I'd guess 20%.

You also get some number of flowers each day. The only truly relevant ones granting +3 familiar weight and +30% meat drops.

It also helps you v-crisis at higher ML with it's navel-like blocks.
I think he may be onto something.


...Maybe.
Except that 1/8*5 = .625 substats/turn, LBoF gives more mainstat than that. And 1/8*20% is only 2.5%, very underwhelming.

And meat drops only help at the nuns, 30% probably wouldn't save a turn.

So we're looking at less than half the stats, 10% less items, all for a chance at 60% of a runaway and an astronomically low chance of a saved turn at the nuns.

I'll take the LBoF.

User avatar
Bantler
Just Plain Old Pie
Posts: 18
Joined: Fri Apr 24, 2009 10:25 am

Post by Bantler » Wed May 06, 2009 8:04 pm

The stats from a firworks are minimal too. Can't you just pull a Bucket if you want consistant item drops so bad? The bouquet could even save a few turns without sacrificing anything.
Calms the moose. You should try it sometime.

User avatar
stupac2
Oh my! Guy with Pie!
Posts: 3027
Joined: Mon Dec 08, 2008 10:04 pm
Location: Stanford, CA
Contact:

Post by stupac2 » Wed May 06, 2009 9:28 pm

Bantler wrote:The stats from a firworks are minimal too. Can't you just pull a Bucket if you want consistant item drops so bad? The bouquet could even save a few turns without sacrificing anything.
A bucket gives .5% items in trade for 1.5 stats. 1.5 stats isn't minimal.

User avatar
Pantsless
Fie the Pie
Posts: 1011
Joined: Fri Aug 15, 2008 4:23 pm

Post by Pantsless » Thu May 07, 2009 2:06 am

lostcalpolydude wrote: On a 2 day SCNP run, fireworks are probably obsolete, though I don't have the experience to say. (Ignoring the crazy run by Pantsless where he did weird things out of necessity.)
While it's true that I had to pull the fireworks in order to get +5lbs for the tower, I'm not certain the jub would otherwise automatically win. In a 2-dayer you're spending a higher % of your turns with a fairy, so:

A) the +12.5% items from the lbof is more important
B) the +1.5 stats adds up more quickly due to more nonsnatch turns

not to mention the other lesser benefits of the fireworks.

I haven't really done the math, but IMHO the fireworks vs jubjub decision isn't crystal clear for a 2-dayer.
[quote="Serra725"]What did you use for the structural part of the bear?[/quote]

User avatar
Bantler
Just Plain Old Pie
Posts: 18
Joined: Fri Apr 24, 2009 10:25 am

Post by Bantler » Thu May 07, 2009 9:56 am

The stats are minimal because you are aren't using that many turns with the fireworks for them to add up. Some of those turns are played on the war when you don't even need stats and would rather have some consistant item drop. Do fireworks flip back and forth every turn between stats and items or is it a random roll? How accurate is the drop rate spaded? I don't know where you're pulling your numbers from but it doesn't matter anyway. You're using a random familiar equipment to gain an insignificant amount of stats and maybe fire a tiny drop bonus in areas where you damn well need items and should use a pumpkin bucket or at least a wash with a mayflower bouquet and more than likely save a few turns.

At the level you're playing at a turn or two matters.

http://koldb.com/searchresults.php?type ... 1year=2007

Jamaica me crazy!
Calms the moose. You should try it sometime.

User avatar
lostcalpolydude
Forum Maestro
Posts: 2940
Joined: Wed Apr 23, 2008 6:54 pm

Post by lostcalpolydude » Thu May 07, 2009 10:19 am

Bantler wrote:Some of those turns are played on the war when you don't even need stats and would rather have some consistant item drop.
This statement is proof that you have a bit to learn. Before the bander, everyone used a sombrero during the war, not a fairy. Now everyone uses the bander. If you're at level 13 when you finish the war, without using clovers/absinthe or at least a stat familiar everywhere possible, you're probably doing something wrong.

Also, the pumpkin bucket is not meaningfully better than fireworks if all you care about is +items during an ascension (+0.5% items is not meaningful, especially compared to the stats). The randomness of fireworks really isn't an issue in this case.
bmaher: Softcore character
lostcalpolydude: Hardcore character (abandoned)

User avatar
Bantler
Just Plain Old Pie
Posts: 18
Joined: Fri Apr 24, 2009 10:25 am

Post by Bantler » Thu May 07, 2009 10:29 am

I'm talking about using the firworks while doing war quests. I'm reading the ascension logs and seeing about 100 turns of non frumious bandersnatch usage. A chunk of those turns must be post level 13 where stats are worth less than items, infrequent meat drops and a few turns.

The average drop rate of the fireworks isn't meaningful. I don't trust the weird math.
Calms the moose. You should try it sometime.

User avatar
lostcalpolydude
Forum Maestro
Posts: 2940
Joined: Wed Apr 23, 2008 6:54 pm

Post by lostcalpolydude » Thu May 07, 2009 10:37 am

Bantler wrote:I'm talking about using the firworks while doing war quests. I'm reading the ascension logs and seeing about 100 turns of non frumious bandersnatch usage. A chunk of those turns must be post level 13 where stats are worth less than items, infrequent meat drops and a few turns.
But they aren't. At least, I've never been at level 13 when I finish the war. On my fastest run, I spent 10k at the legends after killing Wisniewski to reach level 13. Unless you're talking about the turns going up the tower where +items are meaningless.
bmaher: Softcore character
lostcalpolydude: Hardcore character (abandoned)

User avatar
Bantler
Just Plain Old Pie
Posts: 18
Joined: Fri Apr 24, 2009 10:25 am

Post by Bantler » Thu May 07, 2009 10:53 am

How many turns do you use the fireworks after level 13? It looks like about 40 to me. That's 30 stats at the cost of drops only working half the time your entire run and zero chance of saving a few turns. A pumpkin bucket would give better odds for glands and ratchets and a mayflower bouquet would make up for some of the stats.
Calms the moose. You should try it sometime.

User avatar
kindia
Pie in the Sky
Posts: 156
Joined: Thu Aug 28, 2008 2:34 pm

Post by kindia » Thu May 07, 2009 11:03 am

Bantler wrote:I'm talking about using the firworks while doing war quests. I'm reading the ascension logs and seeing about 100 turns of non frumious bandersnatch usage. A chunk of those turns must be post level 13 where stats are worth less than items, infrequent meat drops and a few turns.

The average drop rate of the fireworks isn't meaningful. I don't trust the weird math.
no, I highly doubt anyone uses anything except the stat familiars after the war. lets take a look at your 100 "non-snatch turns" and make a rough estimate of where they come from.

lets see. some of those turns are going to be with a llama trying to get gongs. well be generous and say we farmed 2 gongs day 1 for 10 turns and 3 gongs on days 2 and 3 for a total of 25 turns. so 35 turns to get spleen hits.

now, places where a pixie is more useful than a stat famiiar: harem (5 turns), bat hole (4 turns), maybe 8-bit (putty speeds this up considerably, so 6 turns), goatlet (6 turns), maybe the barrr for rum (10 turns), f'c'le (10 turns) orchard (6 turns), upper chamber (20 turns).

now times to use an oddball familiar. 15 turns in the nuns with a leprechaun. thats it.

now, lets add this up. I hope addition doesnt fall under the category of "weird math" for you.

35 (spleen)
15 (nuns)
05 (harem)
04 (bat hole)
06 (8-bit)
06 (goatlet)
10 (barrr)
10 (fcle)
06 (orchard)
20 (upper chamber)

117 (total)

you will also notice that I was extremely generous with the RNG in multiple places, and decided to leave roflmao completely out of this due to spooky putty. it can easily (and often does) take far more turns to get spleen hits and sometimes to do the upper chamber as well.

tl;dr the only times to NOT use as much +stat as possible is when you absolutely need to have an item dropped, ergo, there is no such thing as "minimal stat gain." fireworks>bucket every day of the week.

EDIT: for the comment that you ninja'ed in before this post: for the glands and ratchets, generally speaking, you are running close to 400% item drops without using the bucket. 400% will guarantee a ratchet drop every time you encounter a tomb rat. any amount of item drops beyond that is erroneous.

dont believe me on item drops being that high? here we go.

100% cyclops
75% sticker sword
75% 35-lb fairy
20% phat loot
20% looting skillz
15% BHH pants
15% flaming shirt
15% KGE eyedrops
15% ice pick
10% gnome skill
5% scrabbler

+365% item drops=modified 93% chance for a ratchet drop every time a rat is encountered.

User avatar
Bantler
Just Plain Old Pie
Posts: 18
Joined: Fri Apr 24, 2009 10:25 am

Post by Bantler » Thu May 07, 2009 12:02 pm

Addition says familiar equipment won't save any turns with stats or items after level 13. Addition says the stat gains from a firebox before level 13 won't save any turns. Weird?
Calms the moose. You should try it sometime.

User avatar
kindia
Pie in the Sky
Posts: 156
Joined: Thu Aug 28, 2008 2:34 pm

Post by kindia » Thu May 07, 2009 2:18 pm

okay, I'm starting to get rather annoyed at you, bantler.
which puts you on a list of about 5 other people including darthdud, so I suppose its good company.

1) if we are talking about usefulness "after level 13" we are talking about having no bearing on the ascension game whatsoever.

2) you talk an awful lot about theories. post some numbers to back up your talk.

3) the battlefield is consistantly one of THE top 5 mainstat gaining areas...look at any log posted on the forums. you would be insane to not use a stat familiar there and get as many stats as possible...the only thing you would need items on the battlefield for is to trade for gauze garters for the shadow, but seriously, you will get so many quarters you wont know what to do with and extra +items is silly.

User avatar
Bantler
Just Plain Old Pie
Posts: 18
Joined: Fri Apr 24, 2009 10:25 am

Post by Bantler » Thu May 07, 2009 2:40 pm

I'm only talking about the firebox being obsolete because of the frumious bandersnatch equipment. There are 3 familiar equipments with 5 pounds and some extras. The fireboxes extras are worth less than the pumpkin bucket or the mayflower bouquet. The latter two are about even and might save a few turns.
Calms the moose. You should try it sometime.

slackerjack
Just Plain Old Pie
Posts: 14
Joined: Wed May 21, 2008 10:27 pm

Post by slackerjack » Thu May 07, 2009 4:17 pm

Bantler wrote:How many turns do you use the fireworks after level 13? It looks like about 40 to me.
If someone's spending 40 turns at level 13 in a softcore run, they're doing something wrong, and they should sort that out before sweating the marginal impact of different secondary-familiar equips.

...or did you mean to say level 12? >.>


As to PPB vs. Fireworks:
There are two reasons why speedrunners don't pull the PPB.
1) It doesn't give bonus substats.
2) It deals damage 30-40% of the time, which interferes with VCrisis.
In return for those setbacks, the PPB offers 13% +item, compared to the Fireworks' 12.5%. The fact that the PPB's item bonus fires every round is basically irrelevant (as long as the Fireworks' 25% bonus doesn't push you over 100% effective drop rate for a needed item--but that should be easy to avoid).


As to Mayflower vs. Fireworks:
This one's a more difficult comparison--basically, you're giving up .875 substat and 10% +item every combat turn you spend with a secondary familiar, in exchange for a chance at an extra freerun or two, and an additional 5-45% +meat bonus at the nuns. The drop rate formula for the flowers was never fully spaded (to my knowledge), which complicates the issue.

For freeruns, remember that the +weight flower isn't enough to guarantee an extra freerun by itself. On day one in particular, if you spend 25 to 48 combat turns with the Bander (I imagine most speedruns will fall in this range), then the +3 lbs. from the flower won't be enough, and you'll need another less-than-5-pound bonus to stack on top of it. You could summon love songs or hearts for a chance at this, at the expense of not trying for a day-one popper.

Anyway, I threw together a very, very loose estimate of the freeruns you might get from this. If I assume:
--3 flower drops on day one and 5 each on days two and three (so 13 total, which may be a bit generous, at least for day 1)
--each flower drop has a 1/5 chance of being the +weight flower
--you use one flower each day, as long as you have any to use
--you have to use the appropriate heart/song each time to reach the next 5-pound threshold
--you summon 9 hearts/songs per day (an arbitrary-ish number; 114 MP)

then I get about 1.74 extra freeruns over 3 days. In addition to the +stat and +item tradeoff for using the Bouquet over the Fireworks, you'll also suffer some reduced chance of summoning one or more poppers, which are themselves more useful than a simple freerun.

As for the potential significance of the +meat, I'll leave that to Eleron's nuns worksheet. I expect the benefit there to be maybe half a turn, at best.

There's much more (and better) math that could be done on this, but so far my intuition is still to take the Fireworks over the Bouquet.

kevbob
Spy vs. Pie
Posts: 98
Joined: Wed Dec 12, 2007 12:32 am

Post by kevbob » Thu May 07, 2009 4:19 pm

everyone knows i don't know jack nor shit about ns13 speed.

and, if i'm looking for any softcore speed info, i'm looking here, on these forums, of this clan.

right?

well, right.

props to you, and ya'll, you know your shit, and you share it, and you are helpful.

and maybe this guy, who is he btw? i never heard of him, is being a little antagonistic,

but,

ease up. don't recite "fireworks fireworks sis boom ba" out of lazy habit. maybe he has a point? maybe it's worth looking at?

seeing as how panstless pulled off a 2 day run the way he did, what's the worst thing a prepped account is going to do if it pullls a mayflower rather than a fireworks? seriously?

just remember your roots. if vlad hadn't been shouted down so much in /hardcore about the hobo, where would you be today? eh?

User avatar
Hoopity
AFH
Posts: 252
Joined: Thu Jul 03, 2008 11:43 am

Post by Hoopity » Thu May 07, 2009 5:35 pm

kevbob wrote:fireworks fireworks sis boom ba
Five times a day!

User avatar
Bantler
Just Plain Old Pie
Posts: 18
Joined: Fri Apr 24, 2009 10:25 am

Post by Bantler » Fri May 08, 2009 12:09 pm

I appreciate the even handed math above. The stat difference between a firebox and mayflower bouquet looks irrelevant. The bouquet might even be able to v-crisis longer to make up?

So the only difference is the 10% average drop rate versus about 2 saved turns. I don't know why you average something that is either on or off. It either works or it doesn't. That is not the definition of average.
Calms the moose. You should try it sometime.

salien
Pie in the Sky
Posts: 115
Joined: Fri Jan 11, 2008 1:50 pm

Post by salien » Fri May 08, 2009 4:54 pm

SJ, your math seems to be assuming you're using the bouquet/fireworks all the time. But all of this discussion is assuming you use the jub-jub on the bander; so you're only using the fireworks/bouquet on non-bander turns, which means very few turns day 1. Maybe around 5-7. So, odds are very low you get more than one flower day 1, and maybe 2-3 on days 2 and 3. So on average you'll get like 1 +weight flower.

Honestly, I do think the difference between fireworks and bouquet is pretty small, because you'll be using it for likely noticeably less than 100 combat turns in a bleeding-edge run. Most likely the 70 or so mainstat from the fireworks will cancel out the 1-2 bouquet runaways, so the extra +10% items is what tips it towards the fireworks. Most of the time you're using a non-bander is when you're item hunting anyway, so that +10% makes more of a difference than it may seem.

slackerjack
Just Plain Old Pie
Posts: 14
Joined: Wed May 21, 2008 10:27 pm

Post by slackerjack » Fri May 08, 2009 9:12 pm

salien wrote:SJ, your math seems to be assuming you're using the bouquet/fireworks all the time. But all of this discussion is assuming you use the jub-jub on the bander; so you're only using the fireworks/bouquet on non-bander turns, which means very few turns day 1. Maybe around 5-7. So, odds are very low you get more than one flower day 1, and maybe 2-3 on days 2 and 3. So on average you'll get like 1 +weight flower.
Nah, I made assumptions based on secondary use only. They just might not be good assumptions. :P

The assumption for number of flowers dropped was influenced by the reports in this thread. Of course, there's not a great amount of data here, which is one reason why my estimate is "very, very loose." :)

Overall, I feel the same as you--there's not a lot of difference between the two, but the extra +item nudges me toward the fireworks. Also, as a practical concern, the LBoF option involves less micromanagement, which is important for people like me who are easily distracted.

Post Reply