|
Re: Some answers about common MT questions... (long) |
Tue, 07 April 2009 16:22 |
|
PaulCr | | Chief Warrant Officer 3 Stars! V.I.P
| Messages: 187
Registered: February 2007 Location: An Island that kinda look... | |
|
I've just done an initial test run giving 12000kt with tech levels from 0 to 155 spread evenly and got a fairly surprising result.
Levels Given
0-59 10
60-71 9
72-83 8
84-95 7
96-107 2
108+ 1
I've done a further test and it doesn't matter how the tech levels are distributed, you can have 26-26-26-0-0-0 or 13-13-13-13-13-13 and you get the same number of levels.
I ran a further few meetings adding some preexisting research to each level and it turns out that it does actually contribute to the next level, I have tested before and the level I got tech in showed at 0 which is why I was sure it was being cleared but I think I remember reading somewhere that existing research can swap fields when gaining a level though a method other than straight research if it causes a switch to the primary research field and maybe that happen the last time I looked in the file data.
[Updated on: Tue, 07 April 2009 16:51] Report message to a moderator
|
|
|
|
Re: Some answers about common MT questions... (long) |
Tue, 07 April 2009 17:21 |
|
PaulCr | | Chief Warrant Officer 3 Stars! V.I.P
| Messages: 187
Registered: February 2007 Location: An Island that kinda look... | |
|
Current
Levels 5000kt 6200kt 7400kt 8600kt 9800kt
0-59 6 7 8 9 10
60-71 5 6 7 8 9
72-83 4 5 6 7 8
84-95 3 4 5 6 7
96-107 2 2 2 2 2
108+ 1 1 1 1 1
6200kt & 7400kt still need checking but it seems pretty clear that those values are almost certainly going to be correct.
Etxra minerals above 5000kt are ignored if you have more than 96 existing tech levels
Report message to a moderator
|
|
|
|
|
|
Re: Some answers about common MT questions... (long) |
Mon, 20 April 2009 06:45 |
|
PaulCr | | Chief Warrant Officer 3 Stars! V.I.P
| Messages: 187
Registered: February 2007 Location: An Island that kinda look... | |
|
Spare resources can definitely seem to switch, in a test bed I've received the 8 fields I expected to get from the MT from the message it gave but the got an extra 1 or 2 on top of that which can only be explained by excess resources being switched between levels, I'm going to do a proper test when I have time to see exactly what happen by giving every field 1 extra resource which won't be enough to actually increase any levels but I will be able to look at what happen after the MT gives it's levels to check the results and deduce exactly how they get moved.
The results should still have a total of 6 excess resources but they'll probably be in the final field with 0 in the others unless a field didn't gain any tech but I also have a sneaking suspicion that there may be an issue with the resource switching which stops the original actually being zeroed so it could end up something like 135246 since in the testbed where I got 2 extra levels I didn't think I would have had enough excess resources to have got both and the other levels I got from the MT also seemed closer to the next than they should have been.
The original test I made did have all excess resources set to 0 and they remained at 0 after meeting the MT so the MT couldn't have been giving partial levels.
Report message to a moderator
|
|
|
|
Re: Some answers about common MT questions... (long) |
Mon, 20 April 2009 13:47 |
|
PaulCr | | Chief Warrant Officer 3 Stars! V.I.P
| Messages: 187
Registered: February 2007 Location: An Island that kinda look... | |
|
m.a@stars wrote on Mon, 20 April 2009 12:28 | I hope you manage to find out all the details on this. As with popgrowth and minerals, this is an area which the Stars! client seems to obfuscate but direct examination of m-file contents should unravel.
|
Just done a very quick test, only 2 runs, all techs set to 0 with 1 resource to Energy, 2 to Weapons etc up to 6 for Bio. Next research field was set to lowest with current field being Prop.
First run got Bio,Bio,Con,En,Weapons all with message that they would continue in Prop, last level was Prop with message that it would continue in Elec. Final resource numbers where 1,2,0,4,8,6 meaning the 3 resources that were in Prop were added to the 5 already in Elec.
Second run got Prop continue in Energy, Energy, Elec, Bio, Bio, Prop all with continue in Weapons. Final resource numbers were 0,6,0,4,5,6 meaning the 3 resources in Prop go transferred to the 1 in energy to make 4, the in Energy then got transferred to the 2 in Weapons to make 6.
It seems pretty clear that any excess resources in a field you are currently researching get transferred to the new field when it is changed, if it changes again then they get transferred to a new field. Presumably he same things happens if you get a level in whatever you are currently researching from battle and switch to a new level.
I'll have to do some test with more complex scenarios since I'm still convinced the previous test I did could not have had enough resources hanging around to complete 2 fields. I did complete 2 levels in different field the year I reached the MT through normal research so that might cause some sort of different result but I can't think how at the moment.
Report message to a moderator
|
|
|
|
|
Re: Some answers about common MT questions... (long) |
Fri, 24 April 2009 18:40 |
|
PaulCr | | Chief Warrant Officer 3 Stars! V.I.P
| Messages: 187
Registered: February 2007 Location: An Island that kinda look... | |
|
There is a blue diamond thingy for changing choosing which object at a location you want to target. There is nothing you can do though to convince the MT to change direction as it is completely random when it does so.
Report message to a moderator
|
|
|
|
Re: Some answers about common MT questions... (long) |
Fri, 24 April 2009 19:29 |
|
PaulCr | | Chief Warrant Officer 3 Stars! V.I.P
| Messages: 187
Registered: February 2007 Location: An Island that kinda look... | |
|
I've just tried it and you can use it to target the MT as a WP1 from elsewhere even if there is another ship there which is what he was asking, targeting it with the fleet that is already there though will not work, it will let you target it but it won't follow the MT which I assume is what you meant, if say you brought 2 separate fleets to the MT and merged them then the merge would have occurred after the MT checked whether they had 5000kt so you might want to follow it with the newly merged fleet, to do so you have to set a ship from elsewhere to target it and then target that with the freighters.
Report message to a moderator
|
|
|
Re: Some answers about common MT questions... (long) |
Sat, 25 April 2009 00:17 |
|
|
IIRC you can target the MT as much as you want for WP1 or higher, the problem occurs when you want to follow the MT when you're already at the same location by selecting it as WP0 - this doesn't work.
Report message to a moderator
|
|
|
|
|
|
|
Re: How often does an MT appear ? |
Sun, 09 January 2011 02:05 |
|
|
Robert wrote on Sun, 31 July 2005 07:23 |
... I did 5 Testgames now and got 5 very different results (years of the MT from 40 to 150)
47/53/57/73/89/149
65/67/109
47/57/59/63/79/89/105/113/117/143
45/55/61/71/95/107/121/134/147
45/50/72/83/99/104/125/129/134/137/139/141
Especially the last game is interesting: 6 MTs in 16 turns!
I also remember a duel where we had 3MTs in 4 turns!
So I truly believe it is completly random...
|
So, in the first 4 of those test games, the MT appeared 27 times in odd numbered years and once in an even numbered year.
In the fifth test game, it was 8 times on odd years and 4 times on even years.
Ok, anything is possible with random numbers, but seriously, that seems very far from random. It could be semi-random but with auto-correlation.
That seems unlikely just from a programming perspective though. The simplest rules would have been x% chance per year after 2440, subject to a maximum of N traders on the board at any time.
But Robert's results are still very far from the center of a binomial distribution. If you flip a coin 40 times and only get 5 tails ... it can happen, but it's not very probable.
Does anyone have better information on this? It seems to me that stats could be generated from the "specified tech level battlesim generator". That program meets a *lot* of M.T.s.
Report message to a moderator
|
|
|
|
Re: Some answers about common MT questions... (long) |
Mon, 07 February 2011 11:43 |
|
BlueTurbit | | Lt. Commander
RIP BlueTurbit died Oct. 20, 2011 | Messages: 835
Registered: October 2002 Location: Heart of Texas | |
|
Random numbers are not a computer problem. The computer only does what the programmer tells it to do. This includes programs as far back as the Win 3.11 days.
I did a quick routine using Python, and generated some random MT years for 5 potential games:
40 67 70 87 92 94 98 99 104 105 116 126 128 134 138 139 150
44 47 57 61 67 69 76 100 107 109 113 117 118 119 123 128
50 53 56 57 59 60 70 74 78 80 81 87 89 93 116 117 124 129 130
50 58 59 61 68 79 82 86 95 102 107 110 113 121 133 140 141
45 51 53 55 58 64 73 81 82 86 87 98 106 109 110 119 121 130 132 136 145
That's random enough to be suitable for a game, and this routine could easily be adjusted to come up with the desired results before using it in the actual game program.
I can easily modify this to come up with more or less MT's, even change the routine to increase the frequency for certain years, whatever I want the results to be.
BlueTurbit Country/RockReport message to a moderator
|
|
|
|
Re: Some answers about common MT questions... (long) |
Mon, 07 February 2011 13:19 |
|
craebild | | Lieutenant | Messages: 568
Registered: December 2003 Location: Copenhagen, Denmark | |
|
As Eagle of Fire says, computers have a problem with generating truly random numbers, so some quasi-random method is used. IIRC modern games use the computer clock to generate a seed number n where 0 <= n < 1 by taking the fractional part of the time in seconds when the game is started on the PC, and then use some quasi-randomization algorithm from there. The question then is how random that algorithm is...
In some games a new seed from the computer clock is used every time a new random number is needed, but that does not work if there are situations where several random numbers are needed with the same number of clock cycles between them every time that situation occurs, as I think is the case in some situations in the battle engine, but using a new seed from the computer clock each time a new turn is generated and then using a quasi-randomization algorithm from there should make the occurrence of the MT effectively random. That might not be the method used in Stars!, though, as there were far fewer CPU clock cycles per second when Stars! was developed than there are now, and asking the computer for the time might not have yielded as many decimals in the seconds back then either.
Med venlig hilsen / Best regards / Mit freundlichen Grüßen
Christian Ræbild / Christian Raebild
Report message to a moderator
|
|
|