Galaxy Legion Forum
http://galaxylegion.com/forum/

Issue with buffs on rift
http://galaxylegion.com/forum/viewtopic.php?f=5&t=44774
Page 1 of 1

Author:  stevo2012 [ Sat Dec 22, 2018 2:37 pm ]
Post subject:  Issue with buffs on rift

Hey all,


so Iv thought Ive noticed this before, but Im sure now - I have activated both the Bio-Con Control and VARA Interface Link abilities on my rift (45x arty) and noted the amount of artifact the planet was producing prior to the activated abilities - 43678 units/hr - this did not change after activating both those abilities. Anyone else experienced this issue?

Current buffs on planet

-Gateway Link (Max) : Production +6%
-Trellith Influence : Artifact +30%
-Solar Radiance : Production +25%
-Gift of the Makers : Production +9%
-Klorvis Supply Conduit : Production +30%
-Regional Science Summit : Research +25%
-Scruuge Logistics : Artifact +20%
-Omni-Harvest : Production +20%
-Q-Pedd Docking : Production +5%
-Lutuma Command Codes : Production +13%
-Bio-Con Control : Production +4%
-VARA Interface Link : Production +4%



Cheers,

Fox

Author:  juiceman [ Sun Dec 23, 2018 9:18 pm ]
Post subject:  Re: Issue with buffs on rift

I dunno man, its glitchy, we had someone in legion with over 102k prod on their rift and those buffs were all working

Author:  Thaumanovic [ Sun Dec 23, 2018 9:36 pm ]
Post subject:  Re: Issue with buffs on rift

Are you sure something else didn't expire as you used them? Two stacked 4% buffs would give about the same as an 8%.

Author:  Deigobene [ Sun Dec 23, 2018 10:10 pm ]
Post subject:  Re: Issue with buffs on rift

It is a known phenomena. There is a level, somewhere over 5x where it seems to round down until it gets over the hump and suddenly stuff will be applied again. It has been this way for years.

viewtopic.php?f=6&t=44730

Author:  Thaumanovic [ Sun Dec 23, 2018 10:13 pm ]
Post subject:  Re: Issue with buffs on rift

Deigobene wrote:
It is a known phenomena. There is a level, somewhere over 5x where it seems to round down until it gets over the hump and suddenly stuff will be applied again. It has been this way for years.

viewtopic.php?f=6&t=44730

People will probably joke about Danmath, but working with floating point numbers can actually be pretty inaccurate.

Author:  FerrusManus [ Mon Dec 24, 2018 2:06 am ]
Post subject:  Re: Issue with buffs on rift

Guide wrote:
Deigobene wrote:
It is a known phenomena. There is a level, somewhere over 5x where it seems to round down until it gets over the hump and suddenly stuff will be applied again. It has been this way for years.

viewtopic.php?f=6&t=44730

People will probably joke about Danmath, but working with floating point numbers can actually be pretty inaccurate.


Even single precision shouldn't have difficulty with numbers in these ranges. A single precision float should be able to be accurate down to about 0.004 (1/256), which would be a 0.4% boost, with numbers as large as 65536, if my calculations are correct. Up to double that, you should still have accuracy down to a 0.8% boost, which I don't think exists. Of course, if they're using something worse than a 32-bit float then that would likely explain it.

Author:  Thaumanovic [ Mon Dec 24, 2018 2:50 am ]
Post subject:  Re: Issue with buffs on rift

FerrusManus wrote:
Guide wrote:
People will probably joke about Danmath, but working with floating point numbers can actually be pretty inaccurate.


Even single precision shouldn't have difficulty with numbers in these ranges. A single precision float should be able to be accurate down to about 0.004 (1/256), which would be a 0.4% boost, with numbers as large as 65536, if my calculations are correct. Up to double that, you should still have accuracy down to a 0.8% boost, which I don't think exists. Of course, if they're using something worse than a 32-bit float then that would likely explain it.


See:
Code:
guide@usebcmathinstead:~$ php -a
Interactive mode enabled

php > echo floor((0.7 + 0.1) * 10);
7

Also, welcome back, FerrusManus.

Author:  FerrusManus [ Mon Dec 24, 2018 4:32 am ]
Post subject:  Re: Issue with buffs on rift

Guide wrote:
FerrusManus wrote:
Guide wrote:
People will probably joke about Danmath, but working with floating point numbers can actually be pretty inaccurate.


Even single precision shouldn't have difficulty with numbers in these ranges. A single precision float should be able to be accurate down to about 0.004 (1/256), which would be a 0.4% boost, with numbers as large as 65536, if my calculations are correct. Up to double that, you should still have accuracy down to a 0.8% boost, which I don't think exists. Of course, if they're using something worse than a 32-bit float then that would likely explain it.


See:
Code:
guide@usebcmathinstead:~$ php -a
Interactive mode enabled

php > echo floor((0.7 + 0.1) * 10);
7

Also, welcome back, FerrusManus.


I guess maybe this is more of an issue for web development then? I don't have the same problem in C++. Switch to fixed point for the win?

Author:  Thaumanovic [ Mon Dec 24, 2018 4:41 am ]
Post subject:  Re: Issue with buffs on rift

FerrusManus wrote:
I guess maybe this is more of an issue for web development then? I don't have the same problem in C++. Switch to fixed point for the win?

There's BCMath, but it relies on you expressing your numbers as strings. The usage is horrible (bcadd("0.1", "0.7"), or bcadd((string) $someInt, "10")) and it scales poorly. PHP offers no other alternatives as far as I'm aware.

Author:  stevo2012 [ Sat Dec 29, 2018 2:13 am ]
Post subject:  Re: Issue with buffs on rift

Thank you guys for the responses.

Buffs didn't expire as I used them, iv noticed this before with these 2 buffs.


I'm really confused about the rest of the responses though lol.

Page 1 of 1 All times are UTC
Powered by phpBB® Forum Software © phpBB Group
https://www.phpbb.com/