Author Topic: Rebuild Behavior  (Read 11553 times)

Offline Extirminator

  • Member
  • Salutes: 56
    • [Rydr]
    • 45 
    • 45
    • 23 
    • View Profile
Rebuild Behavior
« on: June 07, 2014, 11:38:42 am »
Does anyone know how exactly rebuilding works? I mean, full on -

How does the rebuild power of the tool is affecting the number of hits or rebuild power(?) per hit of tool on different components -


I gathered some info here about number of hits for rebuild-

Hull Armor ( Squid ): Mallet - 10 / Spanner - 4 / Wrench - 5

Balloon: Mallet - 24 / Spanner - 10 / Wrench - 12

Light Gun: Mallet - 22 / Spanner - 9 / Wrench - 11

Heavy Gun: Mallet - 36 / Spanner - 15 / Wrench - 18

Light Engine: Mallet - 20 / Spanner - 8 / Wrench - 10

Heavy Engine: Mallet - 30 / Spanner - 12 / Wrench - 15

I have only yet to test out squid armor and I thought I might ask before I go in further to save some work.

Also, if anyone knows too - what is the 'cooldown' for using a tool for rebuild? that is - the time between tool hits upon rebuild.

Thanks in advance.

Offline Sammy B. T.

  • Member
  • Salutes: 154
    • [Duck]
    • 23 
    • 45
    • 45 
    • View Profile
Re: Rebuild Behavior
« Reply #1 on: June 07, 2014, 11:46:16 am »
Here is a data dump I made a bit ago, the spire info may be incorrect now. These were done assuming that a shifting spanner would always be near armor.

Data Dump
A place to keep records of numbers. Probably going to be a mess.

Tool Reference
Rubber Mallet, Repairs 250hp/9s (hp/s 27.7) Rebuild 2
Pipe Wrench, Repairs 120hp/5s (hp/s 24) Rebuild 4
Shifting Spanner, Repairs 40hp/2s (hp/s 20) Rebuild 5

Each swing of a repair tool is roughly .75 of a second. (Actual mean of ten trials of hitting a component ten times was .75508 with a standard deviation of only .02)
      
Various repair tool hits necessary to rebuild an object.
Galleon Hull, 10 Spanner hits                  Min time 7.5
Pyra Hull, 9 Spanner hits *                     Min time 6.75
Junker Hull, 9 Spanner hits                     Min time 6.75
Spire Hull, 7 Spanner hits                     Min time 5.25
Goldfish Hull, 6 Spanner hits                  Min time 4.5
Squid Hull 4 Spanner hits                     Min time 3
Balloon 10 Spanner* or 12 Pipe Wrench hits            Min time 7.5/9
Light Guns, 9 Spanner or 11 Pipe Wrench hits         Min time 6.75/8.25
Heavy Guns 15 Spanner* or 18 Pipe Wrench hits         Min time 11.25/13.5
Turn Engines, 8 Spanner or 10 Pipe Wrench hits         Min time 6/7.5
Primary Engine, 12 Spanner or 15 Pipe Wrench hits      Min time 9/11.25

*The last rebuild swing can be done by switching to a mallet and placing one swing, I will place an asterisk. (useful for making sure the mallet is ready)

Offline Extirminator

  • Member
  • Salutes: 56
    • [Rydr]
    • 45 
    • 45
    • 23 
    • View Profile
Re: Rebuild Behavior
« Reply #2 on: June 07, 2014, 01:10:10 pm »
Here is a data dump I made a bit ago, the spire info may be incorrect now. These were done assuming that a shifting spanner would always be near armor.

Data Dump
A place to keep records of numbers. Probably going to be a mess.

Tool Reference
Rubber Mallet, Repairs 250hp/9s (hp/s 27.7) Rebuild 2
Pipe Wrench, Repairs 120hp/5s (hp/s 24) Rebuild 4
Shifting Spanner, Repairs 40hp/2s (hp/s 20) Rebuild 5

Each swing of a repair tool is roughly .75 of a second. (Actual mean of ten trials of hitting a component ten times was .75508 with a standard deviation of only .02)
      
Various repair tool hits necessary to rebuild an object.
Galleon Hull, 10 Spanner hits                  Min time 7.5
Pyra Hull, 9 Spanner hits *                     Min time 6.75
Junker Hull, 9 Spanner hits                     Min time 6.75
Spire Hull, 7 Spanner hits                     Min time 5.25
Goldfish Hull, 6 Spanner hits                  Min time 4.5
Squid Hull 4 Spanner hits                     Min time 3
Balloon 10 Spanner* or 12 Pipe Wrench hits            Min time 7.5/9
Light Guns, 9 Spanner or 11 Pipe Wrench hits         Min time 6.75/8.25
Heavy Guns 15 Spanner* or 18 Pipe Wrench hits         Min time 11.25/13.5
Turn Engines, 8 Spanner or 10 Pipe Wrench hits         Min time 6/7.5
Primary Engine, 12 Spanner or 15 Pipe Wrench hits      Min time 9/11.25

*The last rebuild swing can be done by switching to a mallet and placing one swing, I will place an asterisk. (useful for making sure the mallet is ready)

As much as I appreciate that, its not what I was looking four, I want the actual modifiers and equations these rebuilds go by.

None of this explains why engines that have more HP than guns, take less time to rebuild. there has got to be some sort of modifier that gets put into account for different components.
Further more, on a squid, 230 HP armor takes 4 swing of spanner. and you can visually see its a quarter exactly each swing. that means, on squid armor, assuming rebuild works by HP of component, it should 'rebuild' 57.5 HP of it's health each swing. But if you compare that to a galleon, 800 HP divided by 57.5 per swing, gives you 14 swings while in reality you need 10. this means that other than a modifier for different components, you have a modifier for different ship's armors too.

I am collecting this data for a program I am writing and I'd rather have just data of the rebuild power, and whatever you are rebuilding to be put in whatever the equations and modifiers in the code work like. So if a developer can check this out for me, it would be very nice :)

Actually that time between swings upon rebuild is helping me though. Thanks :)

Offline Omniraptor

  • Member
  • Salutes: 51
    • [Duck]
    • 27 
    • 45
    • 38 
    • View Profile
Re: Rebuild Behavior
« Reply #3 on: June 07, 2014, 06:15:24 pm »
Remember that rabidly clickety-click-clicking will rebuild or buff noticeably faster than just holding the button down. I consider this a bug, but muse seem pretty indifferent.
Urz even wrote a program to automate rebuilding and make it less rsi-inducing.

Offline Erheller

  • Member
  • Salutes: 9
    • [Duck]
    • 13 
    • 24
    • 14 
    • View Profile
Re: Rebuild Behavior
« Reply #4 on: June 07, 2014, 11:21:30 pm »
I remember awkm saying once that rebuild times for components can be changed independently of each other. This gives me the impression that there is no formula, and component x can be rebuilt in y hits because Muse says so.

I can't find that post, but if I do I'll link it here. It was in the dev app forum for 1.3.3, maybe.

Offline Omniraptor

  • Member
  • Salutes: 51
    • [Duck]
    • 27 
    • 45
    • 38 
    • View Profile
Re: Rebuild Behavior
« Reply #5 on: June 08, 2014, 03:01:58 am »
I'm not actually sure it works that way either. The last rebuild change was the spire's armor, and back then awkm simply lowered the value to get faster rebuilds (there might have been some rounding-down shenanigans). AFAIK it's still a direct function of armor, but the function might not be linear :P In any case PM awkm, he'll be able to help.

Offline Crafeksterty

  • Member
  • Salutes: 73
    • [GwTh]
    • 17 
    • 28
    • 45 
    • View Profile
Re: Rebuild Behavior
« Reply #6 on: June 08, 2014, 05:17:25 am »
Spire Hull, 7 Spanner hits                     Min time 5.25
Goldfish Hull, 6 Spanner hits                  Min time 4.5


There is something incorrect on your research, maybe you just missed it. Spire and Goldfish have the same ammount of armor, thus for the same ammount of rebuild.
Spire is at 6 Spanner hits last time i checked like the goldfish.


Also for the rebuild power research, you can look and count+time the ammount of rebuild power with how many swings one does. If squid is 4 spanner hits (which is 5 rebuild power) to rebuild the hull then it is 20 in total. Making for automatic 5 swings with the wrench, and 10 with the mallet.

If it takes 6 spanner on Spire, that is 30 Rebuild in total etc etc.
« Last Edit: June 08, 2014, 05:21:29 am by Crafeksterty »

Offline N-Sunderland

  • Member
  • Salutes: 281
    • [Duck]
    • 15 
    • 45
    • 23 
    • View Profile
Re: Rebuild Behavior
« Reply #7 on: June 08, 2014, 08:48:10 am »
I'm not actually sure it works that way either. The last rebuild change was the spire's armor, and back then awkm simply lowered the value to get faster rebuilds (there might have been some rounding-down shenanigans). AFAIK it's still a direct function of armor, but the function might not be linear :P In any case PM awkm, he'll be able to help.

Try rebuilding engines and guns, then count the hits it takes. That's enough to tell you that there's no direct relation between health and rebuild time... Except maybe for the hull.

In fact, a bit over a year ago gun and engine rebuild times were increased without their health being changed. I'll see if I can find the patch notes.

Offline Mattilald Anguisad

  • Community Ambassador
  • Salutes: 12
    • [GwTh]
    • 12 
    • 45
    • 30 
    • View Profile
Re: Rebuild Behavior
« Reply #8 on: June 08, 2014, 11:35:45 am »
Q: How does the rebuild power of the tool is affecting the number of hits or rebuild power(?) per hit of tool on different components
A: It's simple every component has a fixed amount of rebuild power required to be rebilt. Divide that by the repair power of the tool (round up) and you get the needed amount of swings to rebuild a component.


As N-Sunderland rebuild power of components isn't related to component's HP (as the component is still badly damaged when rebuilt) - ship armor is the exception and it is related to armor value, but there is a lot of rounding in there, so it's impossible to get exact formula from it.
Major rehaul of rebuild times of components were changed in one of the following patches: 1.1.4 1.1.5 or 1.1.6 - not entirely sure (i belive it was 1.1.4 - I remember it as the great flak nerf and great hawacha op-ness, I could be wrong).

Q: Also, if anyone knows too - what is the 'cooldown' for using a tool for rebuild?
A: Animation cycle when holding down the button seems to be ~1.0 sec, animation cycle when manualy pressing the LMB seems to be ~0,75sec as Sammy B. T. allready wrote.
« Last Edit: June 08, 2014, 11:49:17 am by Mattilald Anguisad »

Offline Extirminator

  • Member
  • Salutes: 56
    • [Rydr]
    • 45 
    • 45
    • 23 
    • View Profile
Re: Rebuild Behavior
« Reply #9 on: June 08, 2014, 06:24:26 pm »
Q: How does the rebuild power of the tool is affecting the number of hits or rebuild power(?) per hit of tool on different components
A: It's simple every component has a fixed amount of rebuild power required to be rebilt. Divide that by the repair power of the tool (round up) and you get the needed amount of swings to rebuild a component.


I have been doing more measurements in practice to get the exact rebuild power value required for each component, thought I'd share the exact values with you. ( if you are wondering it was done with the help of spanner, mallet and wrenches having 5, 2 and 4 rebuild power. That way, with combining them in various combinations, 2 and 5 can get you any value with the exception of 1 and 3 but that doesn't matter because no component has that low of a rebuild value. )

Universal Values -

Balloon - 46
Light Gun - 44
Heavy Gun - 71
Light Engine - 40
Heavy Engine - 59


Armor Values -

Squid - 20
Galleon - 49
Spire - 29
Junker - 44
Pyramidion - 42
Goldfish - 29
Mobula - 39


ship armor is the exception and it is related to armor value, but there is a lot of rounding in there, so it's impossible to get exact formula from it.


Attaching linear relationship of rebuild power and armor HP, it looks quite like a formula to me, BUT, as you said - we can clearly see rounding ups have been done for balance purposes most likely. I did not bother to plot a schematic for the components as they are clearly not connected to HP in any way, shape or form.

I guess that mystery is solved now..

I should also say that on my post at the beginning I said that balloon rebuild with mallet was not 24, it was actually 23 that I meant to write.

Balloon: Mallet - 2423 / Spanner - 10 / Wrench - 12

« Last Edit: June 08, 2014, 06:31:24 pm by Extirminator »

Offline Crafeksterty

  • Member
  • Salutes: 73
    • [GwTh]
    • 17 
    • 28
    • 45 
    • View Profile
Re: Rebuild Behavior
« Reply #10 on: June 09, 2014, 06:25:39 am »



Armor Values -

Squid - 20
Galleon - 49
Spire - 29
Junker - 44
Pyramidion - 42
Goldfish - 29
Mobula - 39


I think your numbers are a bit wrong. By 2 or 1.

The goldfish and a spire reach exactly 50% of rebuild remaining with 3 spanner hits. that is 15 rebuild. Add another 50% witch is 15, accumilating to 30.
I tried 4 spanner hits, then 2 wrenches which is 28. Then it fit just enough for a mallet.
So i tried 4 spanners, and then 4 mallet, then the exact same spot remaining for another mallet before rebuild.
14 mallet swings reaches the exact same ammount of rebuild remaining for a final mallet
(2*15=30)


I didnt test the other ships
Stuff like the baloon is oddly correct. 9 Spanner hits (45) then a sliver of rebuild is requierd left making for half a mallet swing (2). But nothing goes under 2 so i had to 8 spanner, 1 wrench and then mallet to be precicely 46.

same with heavy engine, why is it 59? Just tested and it showed.

And heavy gun, why 71? it took me 13 spanner, 1 wrench , 1 mallet to be precise.
And light gun 44, why not 45?! Would be a buff to the spanner if the numbers were more precise.

The galleon was a hard read, either its 50 or 49.

Wow so if the numbers would go 1 or 2 up or down to precise numbers then the spanner would get a buff.
Baloon should be at 45, light gun 45, heavy gun 70 and heavy engine 60.
For some components it gives the spanner the ability to be 1 whack faster, while the others the wrench a slower choice. Srsly why isnt this a thing?

Offline Mattilald Anguisad

  • Community Ambassador
  • Salutes: 12
    • [GwTh]
    • 12 
    • 45
    • 30 
    • View Profile
Re: Rebuild Behavior
« Reply #11 on: June 09, 2014, 06:59:20 am »
I assume the relative nerf to spanner was an attempt to make gunners more viable (difirence in rebuild power used to be relatively larger).

Offline Queso

  • Muse Games
  • Salutes: 126
    • [Muse]
    • 13
    • 25 
    • View Profile
Re: Rebuild Behavior
« Reply #12 on: June 09, 2014, 11:09:35 am »
Remember that rabidly clickety-click-clicking will rebuild or buff noticeably faster than just holding the button down. I consider this a bug, but muse seem pretty indifferent.
Urz even wrote a program to automate rebuilding and make it less rsi-inducing.

I'm looking into this. It is a bug. Nobody should need hand cramps to win.

Offline Extirminator

  • Member
  • Salutes: 56
    • [Rydr]
    • 45 
    • 45
    • 23 
    • View Profile
Re: Rebuild Behavior
« Reply #13 on: June 09, 2014, 11:29:48 am »
The goldfish and a spire reach exactly 50% of rebuild remaining with 3 spanner hits.

If you are using visual tests you are doing it wrong. it is not exactly 50%, it can never be exactly 50% because of laws of mathematics.

I think your numbers are a bit wrong. By 2 or 1.

I have double and triple tested all of the above values and found them correct*

The goldfish and a spire reach exactly 50% of rebuild remaining with 3 spanner hits. that is 15 rebuild. Add another 50% witch is 15, accumilating to 30.
I tried 4 spanner hits, then 2 wrenches which is 28. Then it fit just enough for a mallet.

You have tested and seen that with 30, the armor is rebuilt, but with 28 it is not. You have not accounted for the possibility of the armor being 29 rebuild power and not 30.


So i tried 4 spanners, and then 4 mallet, then the exact same spot remaining for another mallet before rebuild.

You cannot visually test things. it will just never work out. calculate it.


14 mallet swings reaches the exact same ammount of rebuild remaining for a final mallet

Yes. because 14*2=28 and 5*4( 20 ) +4*2( 8 )=28 ... but yet again you did not account for 29.


same with heavy engine, why is it 59? Just tested and it showed.

Same reasons as I showed above, test it properly.



And heavy gun, why 71? it took me 13 spanner, 1 wrench , 1 mallet to be precise.

13*5=65 | 4+2=6 ----> 65+6=71 ( ? )


And light gun 44, why not 45?! Would be a buff to the spanner if the numbers were more precise.

Again, test it properly. I don't question Muse's decisions, I just calculate.

The galleon was a hard read, either its 50 or 49.

Again, please, don't visualize it, calculate it.

Wow so if the numbers would go 1 or 2 up or down to precise numbers then the spanner would get a buff.
Baloon should be at 45, light gun 45, heavy gun 70 and heavy engine 60.
For some components it gives the spanner the ability to be 1 whack faster, while the others the wrench a slower choice. Srsly why isnt this a thing?

Again, I do not question Muse's decisions I calculate.



A lot of answers for a lot of questions, pardon me.


Note:

*I do not think rebuild power HP of components are necessarily integers, or rebuild power of tool are necessarily integers or another possibility is that the little red circle display is wrong - because you clearly see at some cases it not exactly a half or a quarter or the display.
Again, I cannot confirm this in any way, shape or form with the current data and only a developer with data access could confirm this. But the values I have measured are, I believe, the most accurate we can get for the moment.
« Last Edit: June 09, 2014, 11:32:06 am by Extirminator »

Offline Watchmaker

  • Muse Games
  • Salutes: 55
    • [Muse]
    • 28 
    • 26
    • 17 
    • View Profile
Re: Rebuild Behavior
« Reply #14 on: June 09, 2014, 05:05:25 pm »
The current formula for rebuild is (drumroll)...

HitsToRebuild = ( RebuildBaseHits + (MaxHealth * RebuildHealthMultiplier) ) * PartTypeMultiplier

The "PartTypeMultiplier" is different for each of armor, balloon, guns, and engines. "RebuildBaseHits" and "RebuildHealthMultiplier" are global constants.  "Rebuild power" listed for each tool is the number of "hits" that tool contributes for each swing.

I leave determining the value of each of those constants as an exercise for the sufficiently dedicated reader.