Planetarion Forums

22 Mar 2003, 05:02

I'd like anyone who would like to, to do me a little favor. I'm working on a highly accurate Roid Scan Calculator.

Next time you do a roid scan, record the following:

1.) How many Roids you had before the scans.
2.) How many Amps you had before the scans.
3.) How many scans were performed.
4.) How many Roids were recieved from the scans
5.) How many Amps you have after the scans.

I'd GREATLY appreciate this input. Please eMail me your records at [email protected]

Morden · 22 Mar 2003, 05:24

with the random factor it makes it almost impossible to calc properly

22 Mar 2003, 05:37

I'd still like to try.

So, please, any help is appreciated.

MrL_JaKiri · 22 Mar 2003, 06:08

Isn't this built into PBOT?

zenopus · 22 Mar 2003, 10:01

There is an algorithm, you know.

But since you ask, I have that historic data available from the early phases of this round.
Where the number found differs from the prediction, the prediction is listed in brackets.

Code:

Roids Amps Scans Found
    3   11     4     3
    6   16     3     1 (2)
    7   22     7     5
   12   23     5     3
   15   30     7     5 (4)
   20   36     8     4
   24   55    17    10 (9)
   34   61    10     3 (5)
   37   76    10     4 (6)
   41   93     8     4 (5)
   45  123    16    11
   56  133     9     5 (6)
   61  148     4     2 (3)
                     1 (quest)
   64  165     6     5
   69  160     8     5
   74  180    10     7
   81  193    20    15 (13)
   96  220    20    10 (13)
  106  210    10    6
  112  239    25    13 (15)
  125  260    26    13 (15)
  138  300    36    19 (21)

The number of amps goes down 1 for each roid found, no deviations.

Starbucks · 22 Mar 2003, 11:09

Toop · 22 Mar 2003, 12:13

Quote:

Originally posted by MrL_JaKiri
Isn't this built into PBOT?

PBOT can be used to calc how many roids you will not get...

22 Mar 2003, 17:29

Well... I'll have you know that 2 times out of 5 my calculator was right-on. I just need to tweak it now.

But if none of you want to help me... thats fine.

hyfe · 22 Mar 2003, 17:36

Quote:

Originally posted by WhiteZero
Well... I'll have you know that 2 times out of 5 my calculator was right-on. I just need to tweak it now.

But if none of you want to help me... thats fine.

heh...
Since everything is based on probabilites you cannot estimate the exact number. You can come up with an average number, or a median; but it still won't give you all the information. The final answer to the outcome should be a set of probabilities for the different outcomes; not a single answer.

If you're half decent at mathemathics i can give you the code for my scan calc (see signature) though; just give me a board pm. Its based on the scan formula though and there is little chance actually checking wether that one is correct...

Exact scan calculators have been around since r6 or so i think.. both me and atamur made ones for the strategy forum at the same time (he on a excel sheet; me in php) (others may have made them before too though; but none that i know of)

22 Mar 2003, 17:42

Well, I'd like to do most of the calculations on my own.

So, really, the only thing I need from people are their Roid Scan Stats.

Thanks for the offer tho.

hyfe · 22 Mar 2003, 18:08

Quote:

Originally posted by WhiteZero
Well, I'd like to do most of the calculations on my own.

So, really, the only thing I need from people are their Roid Scan Stats.

Thanks for the offer tho.

Well.. what on earth are you going to do with those stats?

There is a formula for determining success.. taking the results of that formula and reverse engeneering them to something that is bound to more complex than the actual formula in the first place sounds somewhat dubious

And given the number of people who are semi-proficient at maths&stuff at these boards; there is a slight chance you'll get some semi-constructive answers if you explain your brilliant backwards thesis on how to get an exact number for the result of something which really is a distribution of results.

22 Mar 2003, 18:23

heh... I'll keep the details on how my calculator works a secret for now... So right now I just need these stats.

Once I get this thing perfected, I'll give out the source code (it's in Visual Basic).

zenopus · 22 Mar 2003, 19:26

There is no great secret - unless you've managed to code a clairvoyant AI...

The algorithm that gives statistical probabilities for various possible outcomes of a series of scans is well known.

22 Mar 2003, 19:28

Well, I havent recieved any stats yet (email or otherwise). So I guess everyone is just gonna blow me off on this...

thanx

zenopus · 22 Mar 2003, 19:36

Erm, what do you mean you have not received any stats?
I posted a some in this thread...

22 Mar 2003, 19:44

Ah yes, except you. Thanks zenopus!

22 Mar 2003, 19:55

Hmmm... well, according to zenopus' Stats... my calculator was within 1 to 2 roids everytime.

I'd still like some more stats tho. Specificaly ones where the planet has 150+ roids already.

Thanks again!

EDIT: 12 of 22 calculations were only 1 Roid off.
And 7 of the 22 were right on. The remaning were 2-3 roids off.

Starbucks · 22 Mar 2003, 22:55

well, you can do what you intend to do

but you have to do it in assembly, other programming languages are "corrupted"

oh, i forgot, perfect boolean logic will work too

Atamur · 23 Mar 2003, 00:17

Quote:

Originally posted by WhiteZero
I'd like anyone who would like to, to do me a little favor. I'm working on a highly accurate Roid Scan Calculator.

Next time you do a roid scan, record the following:

1.) How many Roids you had before the scans.
2.) How many Amps you had before the scans.
3.) How many scans were performed.
4.) How many Roids were recieved from the scans
5.) How many Amps you have after the scans.

I'd GREATLY appreciate this input. Please eMail me your records at [email protected]

You cannot really produce a calc that's more accurate than the existing calcs which calculate the probability distibution exactly. But if you really want to get ahold of large number of scan results, try bugging Tacticus. Some time ago, he conducted a massive scanning experement to test the scanning formula. He might still have the results somewhere.

SPQR · 23 Mar 2003, 00:18

I don't quite get your intention.

The amount of roids you can get depends on the very simple roids scan formula from the manual and a random chance. Thus you will never be able to create a 'highly accurate Roid Scan Calculator' that tells you the outcome of your roidscans in PA. That would kinda be like telling the future...

SPQR · 23 Mar 2003, 00:19

Quote:

Originally posted by Atamur
You cannot really produce a calc that's more accurate than the existing calcs which calculate the probability distibution exactly. But if you really want to get ahold of large number of scan results, try bugging Tacticus. Some time ago, he conducted a massive scanning experement to test the scanning formula. He might still have the results somewhere.

I think he did that when amps weren't getting destroyed yet tho.

hyfe · 23 Mar 2003, 01:25

Yeah.. that would be correct SPQR..

WhiteZero:
you are aware of the roidscan formula for success?

Code:

0.3 * (amps / roids)

If you really need results; just make a representative selection yourself

.. If you're dodgy at mathemathics; use a a scan calculator with probabalities to make a representative selection.. It isn't that hard coming up with a few numbers

ofc.. if you're planning on proving the scan formula wrong.. that actually is a worthwhile effort; I've never had much trust in pa teams ability to do anything correct :P

23 Mar 2003, 05:49

Well, even if my calc gives the exact same output as many other, web-based calculators, thats fine. My proggie will give people a non-internet based calculator. Just incase the other sites go down, people can have my calculator as a backup... *shrugs*

and I believe that the formula is 30 * (amps/roids)

at0mic.c0w · 23 Mar 2003, 05:55

Quote:

Originally posted by WhiteZero
and I believe that the formula is 30 * (amps/roids)

30% = 0.3

oh and all this business with egting scans from ppl is pointless UNLESS of course u don't believe that they gave us the correct formula...

23 Mar 2003, 06:19

Umm... yes, I'm aware that .3 is 30%... but you dont take 30% * (amps/roids)... it's just 30 * (amps/roids)

ANYWAY! I sure wish people would COOPERATE and take 2 little min on their next roids scan to record that data.
I know... heaven forbid! :eek:

23 Mar 2003, 15:01

Ok, I'm gonna give out a copy of my Calc to a few people so that it can get tested a little more.

Send me an eMail at [email protected] if you want to try it.

hyfe · 23 Mar 2003, 20:32

Quote:

Originally posted by WhiteZero
Umm... yes, I'm aware that .3 is 30%... but you dont take 30% * (amps/roids)... it's just 30 * (amps/roids)

So.. for us who are nitpicks:

you want to use: '30 * (amps/roids)' right?

So.. by using 1 scan at 6 amps, 3 roids.. is my probability to find a roid going to be:

30 * ( 6 / 3 ) = 60 = 6000%
(as you say)
or

30% * (6 / 3) = 0.3 * (6/3) = 0.6 = 60%
(as i believe)

you are funny

23 Mar 2003, 23:10

Well.... using 30* (amps/roids) seems to work perfectly for me.
*shrugs*

hyfe · 24 Mar 2003, 00:01

Quote:

Originally posted by WhiteZero
Well.... using 30* (amps/roids) seems to work perfectly for me.
*shrugs*

that is because you assume the number you're getting is a percentage... and just put a invisible % in the formula instead..

i mean.. you got a number between 0 and 100... what did you think that number represented?

24 Mar 2003, 00:13

Yes... I'm well aware its a %. The outcome is a %. But if you use 30 as a % during the calculation, it's gets majorly messed up.

hyfe · 24 Mar 2003, 04:54

% = 1/100

nothing more, nothing less. If putting in 1/100 in your calculations majorly messes things up, you have a problem

Atamur · 24 Mar 2003, 05:54

This may be the most passionate discussion over percentages that I've ever witnessed.

hyfe · 24 Mar 2003, 06:18

i'm one of those people who have to get the last words

its damn annoying meeting others who're like that too