Home » Stars! 2.6/7 » The Academy » More on the North-South Minefield Bug
More on the North-South Minefield Bug |
Thu, 03 March 2005 19:11 |
|
Crawford | | Civilian | Messages: 3
Registered: September 2004 | |
|
It is known that ships traveling a due north-south course did not trip standard minefields.
The JRC4 patch partially fixed the problem. Ships starting INSIDE the field will detonate mines, even traveling due n-s. However, if the ship starts OUTSIDE the minefield and travels due north-south, it escapes the mines.
But what if the ship isn't going *exactly* north-south, but awfully close to it? After observing a few ships doing just that but not hitting any mines, I did some testbedding.
What I found was that if the ship starts outside the minefield and its final destination is just *slightly* west of due north, it also escapes mine hits. "Slightly" here means one gridpoint, i.e. (1300,1300) to (1299,1400) at warp 10 with perfect safety. It does *not* appear to work if the offset is slightly east.
Has anyone else made the same observation? Or can anyone duplicate these findings in a testbed?
Report message to a moderator
|
|
| | | |
Re: More on the North-South Minefield Bug |
Fri, 04 March 2005 15:52 |
|
Crawford | | Civilian | Messages: 3
Registered: September 2004 | |
|
OK, I've done a bit more testbedding.
In all of the following I'm using stacks of 100 DD's with Trans-Star 10 engines and a single red laser. Ships travel 100 ly at Warp 10. (Stacks are split into 100 separate fleets of 1 ship each, of course.) Minefield is centered at (1313,1190) with radius 107 ly (214 ly diameter).
1) STARTING AT (1300,1080):
This is just outside the minefield, west of due south of center.
1a) To (1299,1180) - one gridpoint west of due northbound
No mine hits out of 100 ships.
1b) To (1298,1180) - two gridpoints west of due northbound
did this twice, with 36 and 41 hits on 100 ships each.
1c) To (1301,1180) - one gridpoint east of due northbound
82 mine hits out of 100 ships.
2) STARTING AT (1313,1080):
This is just outside the minefield, due south of center.
2a) To (1312,1180) - one gridpoint west of due northbound
No mine hits out of 100 ships.
2b) To (1311,1180) - two gridpoints west of due northbound
41 mine hits out of 100 ships.
2c) To (1314,1180) - one gridpoint east of due northbound
80 mine hits out of 100 ships.
3) STARTING AT (1326,1080):
This is just outside the minefield, east of due south of center.
3a) To (1325,1180) - one gridpoint west of due northbound
No mine hits out of 100 ships.
3b) To (1324,1180) - two gridpoints west of due northbound
34 mine hits out of 100 ships.
3c) To (1327,1180) - one gridpoint east of due northbound
75 mine hits out of 100 ships.
ALSO: I *did* check to see whether perhaps the east-west motion happened while the fleet was still outside the minefield. Did so by sending a 101st DD along the same trajectory but at Warp 1 rather than Warp 10, and checking its location every year. Not in every case, and certainly didn't check for 100 years. But in each case the ship was still due north of the starting point even several years into the minefield.
In sum: It appears that (for fleets starting outside the minefield) trajectories slightly to the west of true north can greatly reduce if not eliminate the chance of being struck by a mine.
Report message to a moderator
|
|
| | |
Re: More on the North-South Minefield Bug |
Thu, 17 March 2005 17:10 |
|
Marduk | | Ensign | Messages: 345
Registered: January 2003 Location: Dayton, OH | |
|
From the looks of it, westward movement is done after northward movement but evenly spaced throughout the path. So travel just one point to the west means all the northerly movement occurs, then the fleet moves a single ly to the west - producing a probably non-zero but extremely low chance of a mine hit.
When you move two to the west, you go half the distance to the north, then one to the west, and then the remainder of your northerly travel doesn't register as due north-south. Hence the chance of a minefield hit is roughly half normal... half of your travel is immune and half is not.
I would suspect that (with repeated testing - sample size is important for this sort of thing) moving three to the west gives you a two-thirds normal chance of a mine hit, four to the west gives a three-quarter normal chance of a mine hit, and so on.
My guess is that the code calculates the chance of a hit at each light year you travel east-west, based on the total distance you traveled to get there. So if you are traveling due east through a minefield you get many checks at a very small chance, while movement closer to north-south produces fewer checks at a higher chance each. Aside from the due north-south travel bug, this produces a very similar overall chance of a mine hit for the same speed over the same distance. If this guess is true, I presume the code randomly assigns a location along the hit-test path for the mine hit to occur. That gives you an even-looking distribution of salvage when you chaff-sweep.
I frequently saw this sort of method used for LOS calculations in the early 90s, as I recall it was commonly used in freely available code libraries of the time. The north-south minefield bug may be a consequence of adapting such code from a central origin coordinate system to an 'off-screen' origin. Code intended to deal with the special case of moving across the positive-negative boundary could easily produce weird results when you make 1000,1000 equivalent to what used to be negative coordinates.
Try a similar test going south into a minefield - you may find that eastward motion safer and westward motion provides no benefit.
"The nice guys may not always finish last, but the bad guys do start with certain advantages."
Report message to a moderator
|
|
|
Goto Forum:
Current Time: Mon May 13 01:19:56 EDT 2024
|