Home » Primary Racial Traits » PP » PP permaforming - something is wrong?
Re: correction to spreadsheet |
Wed, 01 March 2006 09:02 |
|
|
Hi!
I ran 4007 hits of 50kT mixed packet today morning and see that permaforming chances are somewhat distorted between different fields (will send results at monday). Could it be that permaforming depends on how far current base value is from the ideal value for PP race (or from the edge for immune field)? Before, I did not see any correlation, but the number of hits was small. In today's test I got most permaforming (96) for gravity, much less less than expected (84) for temperature, and slightly more for Radiation (do not remember exact numbers though). The planet I tested upon is the same as in Excel file (G 45% away, T 18% away and R is 79 - 21% away from the edge).
mazda wrote on Wed, 01 March 2006 15:16 | Some good news.
The value in cells M506, W506 and AA506 should be 0.010475.
This makes the errors for 3 permaforms much smaller.
For M507 it should be 0.00096, in other words about 1 in a 1000 for 4 permaforms.
It's about 1 in 16000 for 5 permaforms !
|
Thanks for correction. I think I have messed up calculation of combination(3 of 10) somehow
I already made automated testing environment that runs 200 hits/minute, so it's not a problem. I'll make a 20000+ hit test to try to get 5 permaforms by leaving it running for the night, and also we will have more reliable results I do not like an idea though to put that huge data into the Excel file for download, so I'll put a summary only.
Quote: | As confirmation of the 300kT capping, in the new theory a 900kT packet should only ever get a maximum of 3 terra in an immune field (except for if it is forced to go to 4 by the permaforming).
As opposed to 1000kT which gets 4 terra reasonably often and never 5.
|
Good point. Will try 900kT to verify what you say.
...
[Updated on: Mon, 06 March 2006 06:24]
WBR, Vlad
Report message to a moderator
|
|
|
Re: correction to spreadsheet |
Wed, 01 March 2006 10:31 |
|
|
Now that everythings automated you should double check that the particular planet's hab, and which field is immune do not contribute to the terra/permaforming chances.
Report message to a moderator
|
|
|
Re: correction to spreadsheet |
Wed, 01 March 2006 10:46 |
|
|
quatch wrote on Wed, 01 March 2006 17:31 | Now that everythings automated you should double check that the particular planet's hab, and which field is immune do not contribute to the terra/permaforming chances.
|
Exactly! It was really hard to reliably write down all the changes by packet manually, so I used the same planet most of the time becuase get used to it. Now I can reliably test alot of different planets.
The only exception:
It's hard to test for low gravity becase 2 several points on the gravity scale for low gravity may have the same gravoity displayed (for example, 0.12g is displayed 2 times for 0% and 1% points on the scale). So when dig up data from the dump file, these would not be distingushed. It means I would not test for low gravity, or test only few cases when 1% of teraforming may happen at most (and this is distinguished in dump file).
WBR, Vlad
Report message to a moderator
|
|
|
Re: correction to spreadsheet |
Mon, 06 March 2006 04:33 |
|
|
Hi!
Strangely enough, I got 5% radiation teraforming by 1000kT packet in 60,000 hits test, and 900kT did 4% max only. So it appears that immune field teraforming is not 75%/300kT. It is not 50%/200kT either - chances distribution is weird in such case, though for 275kT packet they accurate for 0.5/200kT somehow. Looks like it is not a chance/200kT or so, but centered gaussian distribution of (packet_mass-100)/200 points.
Also, now I can confirm few other things asked here:
- teraforming and permaforming does NOT depend on how far current hab value is from the PP race ideal value or from the edge for immune field
- 5% permaforming by 1000kT packet is possible, as well as confirmed both average and distribution for permaforming - it is surely 0.05/100kT
- the same for non-immune teraforming - 0.5/100kT is as accurate as possible
I'm keeping running tests, though this weekend cisvc (indexing service) took control of processor so I got only 6000 hits, that is too small. I'll run tests today and today night so will post final results tomorrow.
[Updated on: Mon, 06 March 2006 06:22]
WBR, Vlad
Report message to a moderator
|
|
| | | |
Re: correction to spreadsheet |
Mon, 06 March 2006 13:10 |
|
|
In the article as "actually impacted" mass was defined as a 2/3 of packet, or, if more exactly, 67% * (1-defence_coverage) * (1-caught by Mass Driver)
WBR, Vlad
Report message to a moderator
|
|
| | |
Re: correction to spreadsheet |
Tue, 07 March 2006 13:42 |
|
|
Hi!
I uploaded new Excel file with results.
http://www.interlogic.com.ua/StarsClub/Downloads/PP_Tera_Tes t_Results.zip (190kB)
Much more to say now
A lot of questions are answered by the last tests:
- Immune field teraforming is whatever you can get as normal teraforming, divide to 2, then if result is non-integer, truncate it down. That explains why 100kT packet cannot do immune field teraforming - 1/2 = 0.5, truncates to 0 always. 300kT packet may do max 3% in normal case, 3/2 = 1.5 = max 1% for immune field.
- Immune field permaforming may be larger than teraforming just because teraforming is halved. So, if 300kT permaforms 3%, there is only 1% of teraforming.
- There is a strict correlation between permaforming and teraforming. It appears that when 100kT chunk chances test is performed, it works just as a simple random number check (from 0 to 1). If number is higher 0.5, we get terraforming. If higher 0.95, we get permaforming. When higher 0.95, it is certainly higher 0.5, so permaforming always goes with terraforming. Because this, all chances are accurate and need no correction (it is not possible to get permaforming without teraforming) - I was wrong before assuming that 0% chances are less than in theory because permaforming. At the same time, as I suspected, there cannot be permaforming without teraforming, except for immune fields because half of terraforming % is used.
- Multiple small packets always do worse than a larger one. It is always better to send 1500kT than 10x150kT. This is the same for permaforming as well as for immune fields.
- It is possible to teraform immune field far behind the teraforming capabilities of the PP race - I managed to terraform immune field (without permaforming) by 40%.
- You cannot teraform normal field behind the teraforming capabilities, though it appears that if permaforming happens, terraforming happens too. It means PP terraforming makes check for 30% limitation AFTER all tests are done. So, if packet does 8% permaforming and 50% of teraformi
...
[Updated on: Tue, 07 March 2006 13:48]
WBR, Vlad
Report message to a moderator
|
|
|
Habitable planet results |
Sun, 12 March 2006 10:54 |
|
|
Hi!
Made habitable worlds terraforming tests this weekend.
http://www.interlogic.com.ua/StarsClub/Downloads/PP_Tera_Tes t_Results_Habitable.zip (139KB)
Conclusions:
- terraforming of habitable worlds works the same as for uninhabited worlds, with exception in case when mass driver is on the target planet
- Defences have no influence on terraforming and permaforming
- IT half of caught minerals rule applies
- No change for AR - world is teraformed the same, even though no people killed
- habitable world teraforming packet size is:
{Arriving packet size} * (1- ({MD level}^2)/({Packet speed}^2)/IF(this is IT, 2, 1))
MD level is level of mass driver at the target planet.
Arriving packet size is what arrives to the planet after all decay during travelling.
To Kotk
Quote: | 3250/3250/3250 packet thrown from driver 6 should hit planet with driver 5 like 994/994/994 by current science (rest is "caught").
|
Confirmed above. No 2/3.
BTW, even though result is more close to 993, Stars! rounds numbers up here. So 993.056 in this example becomes 994.
WBR, Vlad
Report message to a moderator
|
|
|
Goto Forum:
Current Time: Thu May 09 19:45:23 EDT 2024
|