Coordinated spading effort

Muhandes · Post by **Muhandes** » Sun Jun 07, 2009 2:03 pm

yes, Echolocation is +15%. So your -10% was in fact +5%, showing a total of +20% or more. You will need to have even more -item. -15% is not hard to get if you are willing to loose the 40k XP and change talisman.

Jesus · Post by **Jesus** » Mon Jun 08, 2009 1:09 am

If some nice Gadgeteer out there could throw some MD on me, that'd be awesome. It's awkward getting up to +98% when I'm trying to keep my stats lowered. I've hit up HAL (a few days ago, actually) for it, but it appears to be down.

Satan · Post by **Satan** » Mon Jun 08, 2009 1:13 am

Ah, right. Bamse's server administrators said they're not going to allow running of bots due to increase of the load on their servers over the year. So until someone else picks them up, they're down :/

Jesus · Post by **Jesus** » Mon Jun 08, 2009 1:53 am

Yeah, I just read that. Too bad, they're a handy thing to have around.
And I always liked HAL. Such a nice, calm voice...

Muhandes · Post by **Muhandes** » Wed Jun 10, 2009 6:21 am

An update and a thought.

Thought: with enough lassos and red-eyes it might be enough to spade one foes rather than an entire are. For example, lump of carmot seems to have a rather high drop. It would take ~400 of them to spade the foe itself and 500 or so more for each spading project, so I'd say it's still beyond my reach, but maybe a thought for the future.

Update: With a donation of 150 playing cards from TenKPlus I started spading Stacked Deck. So far:

0%: 14/15
+5%: 9/10
+10%: 39/40
+13%: 49/50
+14%: 49/50
+15%: 350/350

Satan · Post by **Satan** » Thu Jun 11, 2009 3:23 pm

Just wanted to note in here, I'll be running the PFE buffbots now, hopefully they'll stay running when I shut down my command line to the SR server. I atleast know they're running right now as it's reading out stuffs. Anyways, Bamse'll let me know if this setup needs adjusting to run 24/7.

Jesus · Post by **Jesus** » Fri Jun 12, 2009 9:35 pm

I'm not sure what I should do next.

Enerbun drop is probably 49 or 50%.

IF it's 49, +98% gives 97.02% chance of drops, while +99% gives 97.51%
IF it's 50%, +98% gives 99% chance of drops, while +99% gives 99.5%

Should I skip to +99%, in which case the very improbable non-drop would confirm 50%?
Or since I've already put 500 turns into +98%, should I keep going with that, with the idea that, if I hit 1000 and still haven't seen a non-drop, 49% is effectively ruled out?

Satan · Post by **Satan** » Sat Jun 13, 2009 5:56 am

Personally, I've always considered 500 to be 'good enough'... But 500 all-drops at 98% should indicate that it's more than 49 and 50, it's 51%. Of course, 97% +items would still give all drops... 96% would put the drop rate at 99.96% if enerbun is 51%, and we don't know if ryme rounds at all on those :/

Muhandes · Post by **Muhandes** » Sat Jun 13, 2009 11:22 am

If you've already put 500 turns at 98% and always got a drop, try getting a no-drop at 97% or 96%.

I'm a bit worried that we made a mistake in one of our previous numbers, so could you please specify how exactly you got 98%? Same for the 97%. All the numbers I used so far except the -20% were pure items, so the numbers were certain.

Jesus · Post by **Jesus** » Sat Jun 13, 2009 11:07 pm

Jesus wrote:I *think* enerbun has a 50% drop rate, or close to it.
To verify, I'm starting with +95% item drop, and I'll work my way up from there.

+95% - 91/92
+96% - 80/81
+97% - 45/46
+98% - 500/500

From page 2

I've kept editing it as I went along.
So it's definitely not more than 50%.

As for what I've been using... At 95%:
20% - hero's cape
30% - age pill
30% - kangaroo
10% - binoculars/isotope meter/gumshoe (I switched these at some point)
5% - nail boots

Switched the binoculars/isotope meter for gumshoe so I could wear more stat reducers.
At 96, I switched out the nail boots for short end of the stick+sticky gloves.
At 97, I switched the sticky gloves for duct tape
While on 98, I ran out of gumshoe, so my set-up was:

20% - hero's cape
30% - age pill
30% - kangaroo
16% - Metal Detector
2% - Pouch of a Couple Lost Things

It took me that long to remember that I could use the Pouch for fine-tuning percentages.

I'll do 500 turns of 99% using this:
20% - hero's cape
30% - age pill
30% - kangaroo
16% - Metal Detector
3% Pouch of Few Lost Things

If something is suspect, I'd guess it was gumshoe. It seems odd that it took so few turns to crack 97%, while 98% is a drop every time... but I could have just gotten lucky (but assuming the drop rate is 50%, +97 gives 98.5% chance of drops... so one-in-46 isn't really *that* ridiculous). But I think it's only possible to get +60-something from pure items, so using other sources is unavoidable.

Muhandes · Post by **Muhandes** » Sat Jun 13, 2009 11:45 pm

Just a quick note, I have to run somewhere.
I don't see the point of doing any more, as you already got 500 drop at X% and nodrop at (X-1)%.
We have to think what this means and maybe start redoing the ones we suspect.

Satan · Post by **Satan** » Sun Jun 14, 2009 12:11 am

Wait, hold up... a non-drop at +97% means it can't be 51%... But +98% at 50% means the final rate is 99%... you should be able to get a no drop at +98%... In fact, it's mathematically impossible not to, since Ryme doesn't use decimals :/ Are you sure you didn't accidentally hit +100%?

Jesus · Post by **Jesus** » Sun Jun 14, 2009 1:05 am

Satan wrote:But +98% at 50% means the final rate is 99%... you should be able to get a no drop at +98%...

I should... but 99% *is* very high, so I'm not that surprised I didn't get one. I'm sure it would happen if I ground away at it long enough, but I think I'm more likely (or quickly/easily) to eliminate 49% statistically than nail 50% definitively. That's the simplest answer, and it points to the same percentage Solitude's brute force spading suggests.

Another 145 turns at 99%, no non-drops (again, further proving 49% unlikely).

Another possibility is that though drop rates are integers, maybe the bonuses from sources other than equipment are not. I doubt that, though. I checked and rechecked my equipment every time I changed something, and fiddled with the pouch at the beginning of each day (the first time, I actually didn't know an equipped partially full pouch remains equipped and fills back up at rollover; I thought it would unequip or stay at the same fullness). So I can't think of any point where I could have been over 100%.

[edit]

Smuggler appears to be... ah... very good.

Smuggler +46% - 77/77
(Age pills + MD, just running out turns of nicely aged here so I can use other stuff)

Smuggler + 36% - 54/54
(Cape + MD)
I'll use this set-up until I run out of MD. I've got it, so why shrug it? The cape really helps extend PP, and since I'm only using the one equipment slot I can bang my stats down further and reduce healing needs.

Jesus · Post by **Jesus** » Sun Jul 19, 2009 11:11 pm

Got item drop info for A renegade robot hive up on the wiki.

Next up... I'm kind of curious what The 'Swift and ...' trilogy drops at... either that or a check of corrupt amulets.

Muhandes · Post by **Muhandes** » Mon Aug 03, 2009 11:52 pm

After a long break from spading to hunt shimmering chaos, I finally finished spading Stacked Deck, and confirmed it is 10%.

As there are no more unspaded item drop modifiers except the smuggler, I'm thinking of two more things we could spade.

1. Non-combat chance. I think it would be very useful to know the non-combat percentage in each area. In some of the areas it's also quite easy to check.

2. Chip drop behavior. Several skills have unknown amounts of item drop bonuses. I'm starting to work on a system that will make it easier to find them, a la the KoL bunnies method.

Post here if you are interested in helping in these or have some idea on anything else you want to spade or thing it is worthwhile to spade.

Jesus · Post by **Jesus** » Tue Aug 04, 2009 10:34 pm

If for some reason someone wanted to verify the drop percentages for items that have a base rate that makes that possible...

The smuggler's box seems to be around 75% (based on 292 encounters with no drop modifiers).

Muhandes · Post by **Muhandes** » Wed Aug 05, 2009 1:48 pm

I started spading chip drop in the neighboring and neighborly neighborhood and I think the data in the wiki is far from accurate.

I have spent ~100 combat turns and this is what I got:

elderly bowler: 6-9
bike courier: 6-10
skateboarding teen: 5-9
frisbee player: 7-11

Later, PKProStudio reported 12 from frisbee player, but no more than that. This leads me to suspect that chip drop data on the wiki was not taken with care and we might need to redo it all.

Jesus · Post by **Jesus** » Wed Aug 05, 2009 9:41 pm

Most of the chip drop data on the wiki comes from before we knew about the effects of the various days; the files I kept on that stuff are from October '07. So you're right.

Since I have about 1000 turns left at the Docks, I might as well keep track of chip drops there too. Although after 113 turns today, the existing numbers look pretty good. Have to remember to take the bribe... I've just been taking the cargo.

Cristiona · Post by **Cristiona** » Mon Aug 10, 2009 8:37 pm

Here's a project for you...

Pentacron flips you a logo to demonstrate. You immediately notice that the logo is nearly identical in size, shape, and feel to Twilight City's plax chips. As you hand the logo back, you wonder if some of the standard chips-detection methods might help you find these more quickly.

How about spading if +chips really does increase logo drops. Since RvW seems to have them as a 100% drop, you'd probably only be able to test in the other two zones, hunting for Aquasson logos. Unless Ford/Paddle aren't 100% before you finish that part of the quest.

Edit: Just thought of a way you might be able to test in RvW: Wear no +chips or +items gear, and use Red Eyes or other effects to reduce item drops and see if you can reduce the drop chance of logos. Then use -chips effects and see if you can get them to not drop.

And, well, hope they aren't 200% drops like URs.

Jesus · Post by **Jesus** » Mon Aug 10, 2009 10:39 pm

Muhandes wrote:2. Chip drop behavior. Several skills have unknown amounts of item drop bonuses. I'm starting to work on a system that will make it easier to find them, a la the KoL bunnies method.

To spade chip drops... wouldn't it be easiest to use the pier security guard? 100 is the max of its chip drop range, and while it might be a bit of a nuisance to verify that you've actually hit the new ceiling, it ought to give an obvious answer.

Although I have some other things to note from the Docks:

On Winsday, taking the bribe got me 92 chips... which adjusts down to 87. Outside of Winsday, the lowest I've seen from it is 90. Also curious is that on Winsday, it never broke the 110 chip ceiling I saw on other days.

I was beginning to think that the 'Swift...' trilogy might be a time-related uncommon because in my ~2500 turn spading, I only got it to drop twice, and that on the same day. Which happened to be a Frayday in which I had two days worth of turns. But I got another today (Sunday) at 8:26:30 PM, so there goes that. If you're curious, that makes the drop rate so far 0.63%. It's anecdotal, but I'm certain item drop modifiers affect that; my main was getting one like every other day when I farmed them. It certainly didn't take me a month to get ten for my collection.

Muhandes · Post by **Muhandes** » Tue Aug 11, 2009 4:48 am

Cristiona wrote:Here's a project for you...

That's an interesting hint. I can say I'm curious, but as you pointed out, this is not an easy task.
Hunting aquasson logos is out of the question. They are not a common drop, and not from an easy to kill foe.
As for the other logos, I'll run a quick test if they are 100% or 200% and see if I can come to a quick conclusion.

Muhandes · Post by **Muhandes** » Tue Aug 11, 2009 5:22 am

Jesus wrote:To spade chip drops... wouldn't it be easiest to use the pier security guard?

Of course this will work, but this is exactly the type of method I am trying to avoid. Let me explain.
From quick testing I did, chips distribution is not uniform. For example, frisbee player, dropping 7-11 chips, does not drop 11 chips 20% of the times, but more likely 5% of the times. So waiting for a 100 chips drop from the guard might be a 1% chance thing, or even less.
Say I don't know if something is 50% or 49%, and I spend 1000 turns in the area and don't get 150. Ignoring the non-combat, there are 5 foes in the area, so only 200 encounters were with the pier guard. The chance of none of this encounters being a 100 chip drop is 13%, so I can't be even remotely certain, and this after spending 1000 turns!

What I'm looking for to spade chip drops is something like the method used in KoL, but I'm afraid I didn't have the time to check if it is valid.
I'll state the method, and maybe someone with more time can check it.

Say something is dropping 7-11 chips. There are only 5 possible drops, and once you got 5 different numbers, you have it all. Now these 5 numbers are sometimes different for each bonus rate. The actual numbers depend on the way chip drops are rounded. For example:
"round" for 49% will give 10,12,13,15,16 while for 50% it will be 11,12,14,15,17.
"floor" for 49% will give 10,11,13,14,16 while for 50% it will be 10,12,13,15,16.
"ceiling" does not differentiate between them, both giving 11,12,14,15,17.

So in many cases, getting all 5 drops, which only takes 20-30 turns is enough to spade something.

This of course assumed chip drops are done the same way as in KoL: decide on drop, then apply bonus, then round, which should be spaded itself.

Muhandes · Post by **Muhandes** » Tue Aug 11, 2009 1:05 pm

I'm afraid my theory was totally wrong.

Frisbee player drops 7-12 chips. That's 6 options. However, with +30% chip drop I got it to drop any number between 9-15, which is 7 options.
road-enraged bike courier drops 6-10 chips. That's 5 options. However, with +30% chip drop I got it to drop any number between 8-13, which is 6 options.

That means that drops are not done as they are in KoL, but some other way. Maybe a number is rolled between 0 and 1, then multiplied by the bonus, and then rounded. Maybe some other way.

However. I might have another idea. It seems that the minimum drop does go by some formula, and minimum drops are quite common.

Take for instance frisbee player, dropping 7-12. With 10% it never drops 7, the minimum is 8. This might also be true with lower bouses. With 30% it never drops 7 or 8, the minimum is 9. So basically, if we can figure out the minimum for which there is no 8 drop, it could be a place to start. I can say it isn't such a straightforward way. For instance, an elderly bowler drops 6-9 usually. With 30% you wouldn't expect her to ever drop 7, but she does. Teen skateboarder drops 5-9 usually, but I never saw a 6 from him with +30% yet. I'll go another day with 30% and report.

Here's my current data. I only did 10% for a very short number of turns so don't consider it complete. In all cases where I state a range I got all numbers in that range.

Code: Select all

                              0%     10%    30%
elderly bowler                6-9    7-10   7-11
road-enraged bike courier     6-10   7-11   8-13
skateboarding teen            5-8    6-9    7-10
superultimate frisbee player  7-12   8-13   9-15

Edit: corrected teen for both 0% and 10%

Cristiona · Post by **Cristiona** » Tue Aug 11, 2009 7:49 pm

Muhandes wrote:Frisbee player drops 7-12 chips. That's 6 options. However, with +30% item drop

Satan · Post by **Satan** » Tue Aug 11, 2009 9:15 pm

Lollll. Anyways, I think it's more likely that Ryme uses a central value with a range. Take 7-12. Your central value is 9.5, range of 2.5 from it. We know that the larger your percentage is, the larger your range gets, so most likely the percentage isn't just multiplied against the central value, it's multiplied against the range as well. Since 10% isn't really complete, we can't make many comparisons between it and 30% to find a correlation, though.

Posting ranges with 0%:

Your neighborhood:
pickpocket - 15-25
big thug - 6-9
thug - 4-7

Neighboring Neighborhood:
skateboarding teen - 5-8
elderly bowler - 6-9
superultimate frisbee player - 7-12
road-enraged bike courier - 6-10

All I've got so far. Did you verify the 9 on skateboarding teen? I spent 26 hours in-game time in neighboring, only ever got 5-8. Although, it did take over 13 to get the 12 on frisbee player, so I suppose it's possible. I'll check some more. Ok, 39 hours and still only 5-8. 52 hours. Still 5-8. I declare the 9 a clerical error. I went all the way to 65 hours just to make sure, it's still 5-8. Damn you person who messed up >.>

Jesus · Post by **Jesus** » Tue Aug 11, 2009 11:11 pm

I was thinking about the pier security guard again, and something jumped out at me because his numbers are so round...

old man from the sea -- 62-103 chips -- 82 XP; chip range=XP -20/+21
pier security guard -- 60-100 chips -- 80 XP; chip range=XP -20/+20
small-time smuggler -- 56-93 chips -- 74 XP; chip range=XP -18/+19
big & tall shoreman -- 53-88 chips -- 70 XP; chip range=XP -17/+18
stevedora -- 51-85 chips -- 68 XP; chip range=XP -17/+17

In other words, the chip drop range is +/-25% of the enemy's XP.
The guard? Based on his 80 XP, we'd expect his chip drops to be 60-100
Old Man? 61.5-102.5
The smuggler? 55.5-92.5
Big & Tall? 52.5-87.5
Stevedora? 51-85

One more example from the wiki:
unidentifiable English king -- 105 XP -- expected range: 78.75-131.25 (it's 79-129 on the wiki)

However, this doesn't work with many encounters simply because everything in their chip drop range is less than their XP. Example: unborn soldier -- 71 XP, 35-56 chips (info from the wiki).

BUT, taking 45.5 as the average of the unborn soldier's range... +/-25% of that gives a range of 34.125-56.875
Taking 9.5 as the frisbee player's average... we'd expect a range of 7.125-11.875, which is what is borne out by Muhandes's data.

I'm not a math guy, so I don't know if there's a better way to put what I'm saying, or if this observation is at all meaningful. All I know is that I'm absolutely certain I don't have to put any more work into spading chip drops at the Docks.

Satan · Post by **Satan** » Tue Aug 11, 2009 11:57 pm

Well. We know that in alot of cases, hp is almost directly correlated to xp, except where Ryme's gone out of his way to make it otherwise (like the GGG). Perhaps it lines up better with the hp, and the xp has some variances (or further variance, rather) that cause the slight discrepancies you see. It also makes a bit of sense for enemies with higher hp to have higher chip amounts (not as much sense as it is for them to have higher xp, but :/).

EDIT: Since hp is known to be variable (I think), maybe that's why some chip values are more rare? It's rarer to get the border hp limits, so it's rarer to get the border chip limits?

Cristiona · Post by **Cristiona** » Tue Aug 11, 2009 11:59 pm

If I were making a game, I'd have everything base off the level of a given foe, save for exceptions (like the aforementioned GGG), so I guess it's not too surprising to see a correlation between HP and chips.

Muhandes · Post by **Muhandes** » Wed Aug 12, 2009 12:09 am

It's a good observation.
it makes perfect sense that like all other things (HP, init, dodge, etc.) chips drops will also have a default average which is based on toughness (ML) with some specific foes having especially higher/lower, same as the HP for GGG, init for flamethrower tower etc.

Keep in mind that this entire effort is not about spading chip drops for foes but about trying to find a way to spade chip drop bonuses on items. Spading the drops from foes is just means to an end.

Muhandes · Post by **Muhandes** » Thu Aug 13, 2009 2:54 am

A day with 20%:

Code: Select all

                              0%     10%    20%   30%
elderly bowler                6-9    7-10   7-11   7-11
road-enraged bike courier     6-10   7-11   8-12   8-13
skateboarding teen            5-8    6-9    6-9    7-10
superultimate frisbee player  7-12   8-13   9-14   9-15

Jesus · Post by **Jesus** » Thu Aug 13, 2009 11:13 pm

Finished up the Docks.

Next up... The Castle. I noticed the other day that the chip drops haven't been really spaded, and given that a few of the drops overlap I'm curious if they vary between enemies. Also, I've been wondering about xentrium ingots.

Jesus · Post by **Jesus** » Thu Sep 03, 2009 10:26 pm

Castle done. Cube next.

I'm interested in Rosencrantz and Guildenstern because their chip drop deviates from ±25% XP. I think the range is the same (±25% XP = a range of XP/2)... I'm guessing it just shifts that 50% more toward the positive end instead of splitting it evenly.

Muhandes · Post by **Muhandes** » Fri Sep 04, 2009 2:29 am

Jesus wrote:Castle done.

Could you say anything about the rate of each of the two non-combats? Is it equal?

Jesus · Post by **Jesus** » Fri Sep 04, 2009 9:58 pm

The data is on the Castle's talk page.

370 instances of Door Number 3 and 397 of Crystal Skulls, so I'm thinking they probably are. Why? Did you have something in mind?

Muhandes · Post by **Muhandes** » Sat Sep 05, 2009 10:43 am

Jesus wrote:The data is on the Castle's talk page.

370 instances of Door Number 3 and 397 of Crystal Skulls, so I'm thinking they probably are. Why? Did you have something in mind?

No, I just didn't read the talk page. I added the data.

Muhandes · Post by **Muhandes** » Wed Sep 09, 2009 1:11 am

I added means for reporting statistics to the wiki. I used this to update the Alchemist page.

What do you think?

Cristiona · Post by **Cristiona** » Wed Sep 09, 2009 1:15 am

How does "{{statrate|274|82|+}}" turn into "(29.9 ± 5.5%)"? Is it 274 adventures with 82 drops?

Jesus · Post by **Jesus** » Wed Sep 09, 2009 8:15 pm

I'll assume that you know what you're doing with the math.

It's certainly more accurate, but I'd question whether presenting it that way really benefits the average user. That's the reason I rounded to whole numbers when I put the stats on their respective pages; sure it's fuzzy, but in general I feel like saying 30% is more helpful than 28.8%. The data is always there for those who want to be more obsessive about it.

When I put it like that, it sounds like I'm busting your chops. Really, I think it's kind of cool, and you've made it very easy... I just don't know that it's necessary.

Jesus · Post by **Jesus** » Fri Sep 11, 2009 10:26 pm

Tomorrow, Jesus will hit level 20.
Back to the Delta, where I'll spend 100 turns with the Smuggler trying to get some idea what ballpark the thing is in.

If someone could hit Solitude up with some Snout, we'll see about finally getting that crossed off the "needs spading" list.

Muhandes · Post by **Muhandes** » Sun Sep 13, 2009 12:08 am

Jesus wrote:I'll assume that you know what you're doing with the math.

It's certainly more accurate, but I'd question whether presenting it that way really benefits the average user. That's the reason I rounded to whole numbers when I put the stats on their respective pages; sure it's fuzzy, but in general I feel like saying 30% is more helpful than 28.8%. The data is always there for those who want to be more obsessive about it.

When I put it like that, it sounds like I'm busting your chops. Really, I think it's kind of cool, and you've made it very easy... I just don't know that it's necessary.

The thing is, I know for a fact that some people take the wiki as a bible, and many don't bother to look at the talk page to find the validity of the data. So if something is, statistically speaking, a wide speculation, I want it to be clear on the page itself, not requiring a visit to the talk page. On the other hand, if we are sure in some a piece of information (and by that I usually mean 95% certainty) it is quite OK to move the statistical information to the talk page and display just the number.

As for 30% being more helpful than 28.8%, I see your point. While 29% is more likely, 28.8% +- 6% is quite likely to be 30% and more intuitive for certain.
How about if I make the template display both? I can do something like

~30% (28.8% +- 6%) *

And clicking the * will point you to a page explaining the statistics. That way one can get an intuitive feeling, but still see that the data is quite uncertain. Or is this too long? Let me know what you think.