Forums › English Language Forums › General › General Discussion

Search

So, you guys are trying to break the game cap DMG?

36 replies [Last post]
Thu, 11/14/2013 - 06:10
Il-Mono-Il's picture
Il-Mono-Il

I've been watching the results of Zeddy and Bopp's testing (and a lot of knights) in the Wiki Editors forum.

The testing with Lost Souls are amazing.
/applauds

But now I'm curious.
What is the cap DMG in Spiral Knights?

If I could make an Loadout just for this, how it would be?

(I've searched for similiar threads, but my slow PC can't complete the search in these forums =X )

Thu, 11/14/2013 - 06:21
#1
Thunder-The-Bright's picture
Thunder-The-Bright
gods ain't gonna help you, son.

max damage should be DA/GF charge on lost soul, with spark revive buff+ frenzied/ deadly cloak buff. I don't know if acheron charge can do the same. the hit must be done with the sword itself, not astral blades /explosions.

Thu, 11/14/2013 - 06:25
#2
Il-Mono-Il's picture
Il-Mono-Il
Yep

I've found some threads, but seems that they're a bit ... old.

@Thunder
About the sparks... how many you need to ... waste ... to get the maximum bonus DMG that they provide?

Thu, 11/14/2013 - 06:45
#3
Thunder-The-Bright's picture
Thunder-The-Bright
gods ain't gonna help you, son.

1. the damage bonus does not scale.

Thu, 11/14/2013 - 06:51
#4
Hexzyle's picture
Hexzyle
When you're this swag this world just bows at your knees

Depends on the data type used to send the damage number.
Seeing as The Ban Stick has to hit for a pretty high number, and the Apocrean takes massive damage when he loses his target, it's probably a short integer or an unsigned short integer, putting the cap at 65535.

Thu, 11/14/2013 - 07:17
#5
Bopp's picture
Bopp
maybe these will help?

"Can you hit for 1000 damage?" (1.5-month-old thread): http://forums.spiralknights.com/en/node/90406

"Overkill?" (front page of Arsenal forum): http://forums.spiralknights.com/en/node/92288

Thu, 11/14/2013 - 07:52
#6
Whimsicality's picture
Whimsicality
Pay no mind to this derp

Well science wins all.

Thu, 11/14/2013 - 07:59
#7
Skayeth's picture
Skayeth
@Whimsicality

Or trinkets..... :/

Thu, 11/14/2013 - 08:19
#8
Krakob's picture
Krakob

Requirements are a Max DI Acheron (or DA and GF, but that's harder), a SoL, and a Maskeraith with Deadly Cloak (if it's more powerful than Frenzied Firestorm, not entirely sure about the numbers). Frenzied Firestorm and Deadly Cloak don't stack. Go to UFSC and find an Howlitzer on the equivalent of D27 if there are Howlitzers there. If not, go for the respawning ones at the end of the D26 equivalent. Die, slap on the cloak, charge your Acheron, and slam a kamikaze Howlitzer with the Charge.

@Thunder The Bright
Lost Souls have no defence but they are not optimal as they can't be found any lower than depth 25. You need Howlitzer Heads, which share the defence value 0 with Lost Souls but can be found in the lowest of depths in UFSC.

@Anyone wondering what Hexzyle is talking about
See:
http://en.wikipedia.org/wiki/Short_integer

Thu, 11/14/2013 - 08:34
#9
Zeddy's picture
Zeddy
@Krakob

Screw your Howlitzer. Go straight to Darkfire Vana and use a recon flare to drop his defence to 0.

Thu, 11/14/2013 - 08:51
#10
Krakob's picture
Krakob

Oh, right. Forgot about that. Kids these days with their barriers and flares and Obsidian weapons and phones with monitors and earphones and skateboards and caps and abbreviations and and and and and and and and and so on

Thu, 11/14/2013 - 10:38
#11
Rangerwillx's picture
Rangerwillx

My idea for the most damage possible was GF charge (now acheron charge I guess) + max dmg + SoL + disentegration ray/recon flare + deadly shadow cloak. It'd be versus a green mini-slime at a D28 danger room.
then again, the green slimes probably already have no defense.

Thu, 11/14/2013 - 10:55
#12
Krakob's picture
Krakob
@Anti Dagger

Green Slimes have defence. A small amount but still some. Howlitzer Heads and death marked enemies have no defence, on the other hand. Do note that Disintegration Ray would fill no purpose if the enemy is already at 0 defence. UFSC is lower than D28, also.

Thu, 11/14/2013 - 12:38
#13
Momofuku's picture
Momofuku
@Zeddy

I have been unable to deathmark Vanaduke or Royal Jelly King with recon flares so I doubt it would work on Darkfire Vana either.

I could be wrong, but I don't think the death mark works on bosses. Which was the only use for saving up those flares, really.

Thu, 11/14/2013 - 12:47
#14
Fangel's picture
Fangel
You can death mark bosses

I'm not entirely sure how large the recon flare's radius is, but if you plop a recon flare right next to any enemy, bosses included, they get death marked. I'm unsure if they need to remain in the radius for the mark to apply after the warning is sent or not, but they can be marked.

Thu, 11/14/2013 - 12:59
#15
Krakob's picture
Krakob
@Momofuku

No need, Vana spawns other mobs that you can mark.

Thu, 11/14/2013 - 13:05
#16
Flowchart's picture
Flowchart

SK is written in Java which doesn't have unsigned integers. They could be using short (max 32767) but most likely they are using plain ints with possibly a hard coded cap.

Also, all floors in SLs seem to be the same depth as damage levels are the same. So you could just try the first floor of one.

Thu, 11/14/2013 - 13:16
#17
Skayeth's picture
Skayeth
@Everyone

Screw everything, max UVs. max trinkets, snarby armor, acheron, its spamming time.

Thu, 11/14/2013 - 15:50
#18
Klipik-Forum's picture
Klipik-Forum
I never thought I would get the chance to use this :o

@Annilation: As you can see in my highly unfinished and bad Chart of UV Science, that's like, entirely pointless. You get the equivalent of two full max damage bonuses, otherwise known as a colossal waste of money and equipment space.

Thu, 11/14/2013 - 16:30
#19
Zeddy's picture
Zeddy
@Momofuku

I've deathmarked Vanaduke before. The timing for it is tricky.

Thu, 11/14/2013 - 17:52
#20
Il-Mono-Il's picture
Il-Mono-Il
lol

Bopp, Zeddy and Krakob:

Thx a lot!

Sorry Bopp, i'll never forgive myself... If it was a snake =x

Thu, 11/14/2013 - 19:49
#21
Rezzler's picture
Rezzler
Its not hard to deathmark

Its not hard to deathmark vana, if you dont mind getting scalded for a few seconds.

But whats a third-degree burn compared to winning 3 tokens? Yeah, NOTHING. HA

In other news, 4* weapon with med dmg bonus beats 5* weapon. Game.

Thu, 11/14/2013 - 19:54
#22
Skayeth's picture
Skayeth
@Rezzler

Deathmarking vanaduke takes skill.
Especially when one is trying to take charge
When one player takes up his fill
Then it its time for a deathmark.

Thu, 11/14/2013 - 21:16
#23
Xtweeterx's picture
Xtweeterx
2 words

Blitz-Strafe

Thu, 11/14/2013 - 21:55
#24
Skayeth's picture
Skayeth
@Xtweeterx

Once I had a party full of Blitzers. Killed him in no time AT ALL.

Fri, 11/15/2013 - 00:42
#25
Hexzyle's picture
Hexzyle
When you're this swag this world just bows at your knees

They could be using short (max 32767) but most likely they are using plain ints with possibly a hard coded cap.

I just said that. What do you think short means? It stands for Short Integer

Fri, 11/15/2013 - 01:45
#26
Il-Mono-Il's picture
Il-Mono-Il
@all

Besides that lucky guy (the one who achieved 1k damage), no one break his record?

Fri, 11/15/2013 - 02:32
#27
Flowchart's picture
Flowchart

@Hexzyle, you said short integer OR unsigned short integer. I already said unsigned does not exist in Java, thus it must be short integer or regular integer.

Fri, 11/15/2013 - 03:01
#28
Hexzyle's picture
Hexzyle
When you're this swag this world just bows at your knees

There's no such thing as signs on a binary level. 0000001111101000 is one thousand written in two bytes of data. When you read this as a signed short, all you're doing is changing the number that you start counting from, from 0 to -32767. 0000001111101000 is still "One thousand more than the base" regardless of what "data type" you're reading it as: As an unsigned short, that'd show as 1000, as a signed short, it'd show as -31767, as a boolean it'd show as 0 (Since we start at 0, 1 means an odd number)

TL;DR
The perceived value of a number can change depending on what format you read it in. Doesn't change the hard cap of the number (the distance between minimum value and maximum value) which, for a two byte number (the size of both a short and an unsigned short) is 1111111111111111 (65535. A byte is made up of 8 bits. Each bit is either a 0 or a 1. For a two byte number, that's 16 bits, as you can see.)

unsigned_short = (signed_short + 32767)

There's your unsigned short.

I strongly doubt they're using plain integers, that's double the message byte size that they could be using, that'd be inefficient when it comes to bandwidth usage. I mean, since when do we need to do damage above several million?

Fri, 11/15/2013 - 03:09
#29
Flowchart's picture
Flowchart

It doesn't matter what it's stored as internally, since they are using a high level language they are limited to what the language lets them use if they're sticking to the predefined data types. Unless they want to get every number they use, check if it's negative, and add each time for each calculation. Yet somehow that's more plausible to you than simply using integers to begin with.

Fri, 11/15/2013 - 04:27
#30
Hexzyle's picture
Hexzyle
When you're this swag this world just bows at your knees

1. You're assuming that the server is written in java. We don't know anything of how the server handles numbers.
2. Programming language isn't what gets sent through the network. Packets, constructed of bytes, are. This is a crucial point as the damage numbers we see are a result of calculations performed server-side, and the result is simply sent to the client as a number.

Unless they want to get every number they use, check if it's negative

Why would you need to check if it's negative? Adding 32767 to a number from a data type that doesn't support numbers below -32767 causes the result to be positive, regardless.

add each time for each calculation

Except the fact that next to nothing regarding combat is calculated on the client, save for a few simulated animations. Apart from the damage owner and target (which are handled as seperate parts of the packet, anyway) all the client has to do is
damage = receive(number)
print "damage" at target.x, target.y

To that first line put a +32767 if the message was sent as an unsigned short.
Sure, if the developers didn't think they needed those extra 32767 values, they could have started "0 damage" at 0 on the short spectrum (32767 on the unsigned short spectrum) and perhaps use a negative damage value as a replacement for a "heal" message.

Fri, 11/15/2013 - 05:51
#31
Flowchart's picture
Flowchart

The point of your last post was that signed and unsigned had the same range, and they were represented with the same amount of bits, they were identical, and thus I said the same thing you did, even though I didn't. One is natively supported, the other isn't.

I'm not sure why you think it's more likely for them to take a number from 0 to 65535, subtract 32767 (actually, 32768) from it, send it to the client that way so it has to readd it than to simply use a 32 bit value to begin with.

You are assuming complexity for the sake of argument. Regardless of the language the server uses, why would they not send a structure that is compatible with the client?

Fri, 11/15/2013 - 05:56
#32
Hexzyle's picture
Hexzyle
When you're this swag this world just bows at your knees

You have a point. I tend to overthink things when I plan out how I'm going to use a data type. I don't think many others program like me.

Fri, 11/15/2013 - 07:49
#33
Il-Mono-Il's picture
Il-Mono-Il
OMG

Can we back to the main question?

XD

Fri, 11/15/2013 - 07:51
#34
Skayeth's picture
Skayeth
@Monocelha

Agreed.

Fri, 11/15/2013 - 14:58
#35
Zeddy's picture
Zeddy
@Hexzyle

"I strongly doubt they're using plain integers, that's double the message byte size that they could be using, that'd be inefficient when it comes to bandwidth usage."

We're both talking about the game with ten-year old graphics running far slower than most modern high-definition games here, right?

The game that stutters for a good second or so when I'm opening a menu?

The Java game that runs less efficient than most 3D javascript games I've tried?

Wouldn't surprise me they were sending damage numbers as doubles.

Sat, 11/16/2013 - 02:09
#36
Hexzyle's picture
Hexzyle
When you're this swag this world just bows at your knees

Yes, I was fully expecting someone to comment on my "efficiency of network traffic" comment XD

Powered by Drupal, an open source content management system