Pages: 1 [2] 3 :: one page |
|
Author |
Thread Statistics | Show CCP posts - 0 post(s) |

Equium Duo
Minmatar Gauss Codazzi
|
Posted - 2009.02.14 12:57:00 -
[31]
Edited by: Equium Duo on 14/02/2009 12:57:22 Guys please, lets just go back to what we love, the way stargate did it!
You need 6 probes! To for 3- intersecting lines, then your ship is the point of origin!!
Obvious!!
Sheesh..
This is how i think this is working,
You have two intersecting lines AB and CD, with a scan deviation for them so the way i uderstand it you get an overlapping area in the middle as your scan result,
Here is a picture of how i imagine it works (it's 2d, but i think it applies to 3d aswell)
|

Hugh Ruka
Exploratio et Industria Morispatia
|
Posted - 2009.02.14 13:08:00 -
[32]
Originally by: Kopkiller Edited by: Kopkiller on 14/02/2009 12:28:20
Originally by: Hugh Ruka
Originally by: Ayari The way I thought triangulation worked was that the two points of reference report the direction of the target, but not the range, the range to target is calculated by the intersection of two directions.
This works in 3d space too, because the direction lines will always intersect at the right point. Obviously the new probe method doesn't use real triangulation.
as was said, you need 3 reference points to calculate position in 2d space. however you are scanning in 3d in EVE.
Stop saying stupid things. Triangulation works by its principle indepedently of number of dimensions, or you just didn't understand how it works at school.
And it's not the problem anyway, 99% of eve is not realistic "real life" wise.
forget the word and think ... think 3D
A plane is definded by 2 vectors, means you need 3 points minimum.
to define a line in 3d, you intersect 2 planes - 6 points
to define a point, 3 planes - 9 points
now you only have your probe and a distance measure, means each probe forms gives a spherical reading to the same signature.
intersecting 3 spheres does not give you a point reading, it gives 2 points (or a line if you want). 3 circles in 2D give a point, however you are missing one dimension in 3D.
SO YOU NEED 4 probes ... --- SIG --- CSM: your support is needed ! |

Space Wanderer
|
Posted - 2009.02.14 13:42:00 -
[33]
Edited by: Space Wanderer on 14/02/2009 13:42:26
Originally by: achoura Tri means three, .... Q. Since you only need three, but are launching four,
*groan*... another smart one who thinks he knows math. Man, the info in the dev posts are correct, the mechanism by which probes work (distance detection) require four probes minimum. If you, or anybody else, think three should be enough, you either did not read the dev posts or you are just not material for a prober.
No matter to me, devs have set up what, with some tweaks, might be a very good system. Kudos to them, and I hope they manage to fix the issues with scanning down ships. I wonder, to this aim, whether the directional scanner is now obsolete. Too easy to spot incoming hostiles/probes, and get away, by using it.
|

mamolian
Cruoris Seraphim
|
Posted - 2009.02.14 14:02:00 -
[34]
Originally by: Deva Blackfire
Triangulation (with 3 probes) works ONLY one one plane. If you add "depth" (or 3rd dimension) you need 4th point (probe) to get exact location.
This 
-----------
|

Haral Reimo
|
Posted - 2009.02.14 14:06:00 -
[35]
Originally by: Joseph Juneau Yeah as previusly stated in 3d space if the info from 1 probe is only distance u need 4 probes (check how gps works, where u need at least 4 satelites in fov to be able to get ur location)
Not quite. You only need 3 satellites with GPS, because although that gives you two points (see below for why) you can discard one as it will be in orbit above the GPS satellites and they only broadcast downwards... It's actually a little more complicated with GPS as you don't know the exact distance (unless you have a synchronised atomic clock with you), you only know the difference in distance between you and each satellite
Originally by: Bimjo how probes work :
1 probe : scan result is a sphere around probe , wormhole can be anywhere within, reason is one probe hasn't got enough signal strength to pinpoint location.
Actually, the scan result is only the surface of the sphere, the scan point cannot be within. The reason is that the probe knows how far away the point is, but not the direction.
Originally by: Bimjo
2 probes : if their result spheres intersect we get the RRR "red result ring"
The intersection of two hollow spheres is a circle. The point will be somewhere on that circle as they are the only points that are the correct distance from both probes.
Originally by: Bimjo
3 probes : if all three result spheres intersect we get a point in space,and 3 probes are still not strong enough to pinpoint the result accurately,remember a 500k deviation means we will be warping to the wrong grid !!!
Not true. A 3rd sphere will intersect the circle in two places
Originally by: Bimjo
4 probes intersecting : result[8)
The 4th sphere will only intersect one of the two points, identifying it as the correct location
|

Kazang
Gallente Arbitrary Freedom
|
Posted - 2009.02.14 15:01:00 -
[36]
Edited by: Kazang on 14/02/2009 15:03:33
Originally by: Hoshi Seems pointless to discuss triangulation considering that the scan system doesn't use triangulation in any form it uses Trilateration.
Quote: Trilateration is a method for determining the intersections of three sphere surfaces given the centers and radii of the three spheres.
3 not 4.
Use 4 probes rather than 3 is just stupid, for both balance reasons (have you tried to move probes around when you have 4 or more probes on top of each other ) and plain maths reasons.
It seriously makes me doubt the intelligence of the dev team, like the typo of "analize", and thats just depressing 
Anyone who knows anything about trilatertion is doing a major facepalm anytime they open up the scan UI. Even the fact the they called it "triangulation" is deeply confusing, triangulation uses angles (the clue is in the name ) to find distances. It does not use distances to find postions.
Even without high horse "ive been to uni and know what i'm taking about" bull**** chest beating from me, its not an intuitive or effective system to use in game. For lots of reasons: the aforementioned moving probes once they get very close together, the whole cant stack used probes thing, the lag that will be caused by every player in EVE having 4 or more probes out when this goes live on TQ. And probably more that haven't occurred to me yet.
Kazang
|

achoura
|
Posted - 2009.02.14 15:25:00 -
[37]
Originally by: Space Wanderer *groan*... another smart one who thinks he knows math. Man, the info in the dev posts are correct, the mechanism by which probes work (distance detection) require four probes minimum. If you, or anybody else, think three should be enough, you either did not read the dev posts or you are just not material for a prober.
No matter to me, devs have set up what, with some tweaks, might be a very good system. Kudos to them, and I hope they manage to fix the issues with scanning down ships. I wonder, to this aim, whether the directional scanner is now obsolete. Too easy to spot incoming hostiles/probes, and get away, by using it.
..and if you were as smart as you think you are you would have noticed i already said the server knows the final location in advance and is not calculating anything. Given that the ui is so bloody awful (i.e. it hinders more than it helps) and that it take so long (making it worthless for pvp) the few seconds saved by not using probe no. four i.e. not having to attempt to put it over the first three probes without moving them, adjusting their size, the camera, or their placement, balances out. If the theoretical few seconds required to move probe no. four matters so such add it to scan time.
There's simply no point is using the final probe for differentiating the locations when the server can do that for you while making the actual probe placement less frustrating and having one less piece of data floating in space. Currently in busy 0.0 systems just dropping one probe in a busy spot finds ssed ships, in future it will require four (preferably three) which will have to be relayed to every client in system. The less data needing shifted the smoother the game runs, and as the people behind the "need for speed" know, afew little saving add up.
For the record, using the distance from four points to calculate a fifth is not triangulation, triangulation uses angles. As i already said, i find it incredible that someone would actually label it thus in the notes, either they are doing the math and don't know the manes or are writing notes for something they have no clue about. ***The EVE servers and their patches*** |

mamolian
Cruoris Seraphim
|
Posted - 2009.02.14 15:50:00 -
[38]
Originally by: Haral Reimo Stuff
Correct. Harals post explained visually:
Pic 1 Pic 2 Pic 3 Pic 4 Pic 5
-----------
|

Nyphur
Pillowsoft
|
Posted - 2009.02.14 15:50:00 -
[39]
Originally by: Space Wanderer now probes report only a scalar value, i.e. the distance of the signature. If with this bit of information you can't understand why you need at least four probes to locate a spot
Quadlateration. I wrote a program to do this given co-ordinates from 4 known scan points and their distances from the target back when probes weren't out. Then CCP made probes and the thing was pretty useless :S. Still nice in theory and the maths for it is very fast if you know how to do it, is the new system actually using that?
|

Space Wanderer
|
Posted - 2009.02.14 16:16:00 -
[40]
Originally by: achoura There's simply no point is using the final probe for differentiating the locations when the server can do that for you while making the actual probe placement less frustrating and having one less piece of data floating in space.
So you are saying that in your opinion using four probes is not good game balancing. That might be true. But don't hide behind pseudomath to say that. You are asking devs to shortcut a perfectly working, and correct, system for sake of what you believe to be increased playability. That is fine. Or you may try to "prove" that devs math is wrong. That is silly.
|
|

BiggestT
Caldari Resurrection Skunk-Works
|
Posted - 2009.02.14 16:21:00 -
[41]
Slightly off topic but..
I find it hilarious that people claim that triangulation is useless in a 3d world (with depth), when thye use it themselves in a world of depth (the real world ofc).
Your using it right now!
Your eyes use triangulation to judge depth, distance and an overall location in a 3-D world.
As to why probes cannot use the same function I have no idea..seems very simple to implement... EVE history
t2 precisions |

achoura
|
Posted - 2009.02.14 16:42:00 -
[42]
Originally by: Space Wanderer
So you are saying that in your opinion using four probes is not good game balancing. That might be true. But don't hide behind pseudomath to say that. You are asking devs to shortcut a perfectly working, and correct, system for sake of what you believe to be increased playability. That is fine. Or you may try to "prove" that devs math is wrong. That is silly.
There is no need to prove anything, triangulation uses angles to calculate location, this is a simple fact which any standard grade math student (read 14/15 year old) can tell you.
Though yes, manipulating there overlapping probes in the map is far, far easier than attempting four. ***The EVE servers and their patches*** |

Kazang
Gallente Arbitrary Freedom
|
Posted - 2009.02.14 16:46:00 -
[43]
Originally by: mamolian
Correct. Harals post explained visually:
Pic 1 Pic 2 Pic 3 Pic 4 Pic 5
No your image 4 is incorrect, your thinking the 3rd probe result is circle or a flat plane, it is not a plane. It is a single linear distance, which is visualy represented as a sphere, the probes give a radius as result. 3 spheres can only intersect at a single point in your image, the distance and 3rd sphere is tangential to the result of the other 2 probes, which is represented as a circle. In the scanner UI this a red circle. You are not thinking in 3D.
Plan and elevation view of the probes
A,B and C are the probes, the black lines and circles represent the scan result. Its a single distance. Which expands to a sphere. The red line in elevation is a radial result from 2 probes, it is a circle in plan. The 3rd probe's result will be a sphere tangential to the red cirle, giving a single result! The yellow and blue lines are the results of the 2 other probes with respect to each other.
However, a problem does arise if the origin of the 3rd probe is within the radius of the other 2 spheres, thus possibly giving 2 positives like so.... one of which is false. But it is quite difficult to achieve this result, 99% of the time it will work fine with 3 and you will get a single result, like so. Using 4 probes does somewhat remove this problem but it still possible get a false positive with 4 probes, just even more unlikely than it is with 3. Its a lag and frustration causing, ultimately inelegant way of solving the problem. Kinda like coding in python
The simple way to stop false positives is to simply place the 3rd probe far enough away from the other 2 to get an accurate result. Not to add more probes. Complicated maths will often end up with multiple answers, only one of which is correct, its up to the person doing the calculation to use his/her logic to determine the correct one.
And this kids is why computers crash.
Kazang
|

COMMANDER KATHRYN
Gallente DEATHFUNK Doctrine.
|
Posted - 2009.02.14 16:48:00 -
[44]
aaack!! leave it to ccp to (TRY, FAIL )fix what wasnt broken. The new scanning interface is way to clumsy why not just have the 4 drones automatically quad around a specific(player chosen) spot. As you increase the scan range the drones move outward from each other automatically. Instead we have to fight with moving them individually which sucks as I keep trying to change the scan range (by accident)while clicking the arrows to move the drone. It really becomes an issue when there are several signatures in close proximity to each other.
But what should we expect while they try to change to much with in the game in one upgrade.
At this point the whole of the new scanning/probing interface seems a bit to rushed
|

Caius Severus
|
Posted - 2009.02.14 16:51:00 -
[45]
Why are people having such trouble understanding the need for 4 probes? Forget the fact that the server knows where the target is, that is irrelavent. Your ship is not considered one of the locations used to measure distance, only the probes are.
It has been said that these new probes determine a distance to a target and nothing else, with an increasing margin of error according to their range. This gives the following effect:
1 probe is sufficient to determine the distance of a target, placing it anywhere on the surface of a sphere that size.
2 probes is sufficient to determine the location of the target anywhere on a circle where the surfaces of the spheres from each probes intersect.
3 probes narrows it down to 2 possible locations.
4 probes eliminates the incorrect one of the previous 2 locations giving a final location of the target which you can warp to. If the probes are too far away the signal is poor and you need to move them closer to get a warp in.
Incidentally, this is exactly the same method a GPS receiver calculates its location - by determining the distance to each satellite. If you use one, you will notice that you can get a 2d fix with 3 satellites, because the 2nd incorrect location can be rejected because the reciever knows that it has to be approximately on the surface of the earth, and the wrong one will be off in space somewhere. A 4th satellite allows a positive 3d fix and gives altitude as well.
|

Sophia Truthspeaker
THE INTERNET. Goodfellas.
|
Posted - 2009.02.14 17:07:00 -
[46]
Concerning triangulation. I understand traingulation as having to known points, and knwoing the angle at which the third point is. Then you know the position of the third point. Pretty easy. The fun part is, it doesn't matter if you are in a 2d, 3d or 86d environment. A triangle will always be on a plane. If a chair has three legs, all three legs will touch the ground. There is no case in which one leg of the chair doesn't reach the floor unlike a chair with four legs.
That being said, it isn't triangulation used with scanning. It measures (like gps) the distances between points. If you got one probe and you know how far away your other point is, it is in a ball around the prob. If you got two probes, you get a circle, because a circle is the intersection between two balls. Three probes, two points, because the intersection of three balls are on two points. That is the reason why you need 4 probes.
_________ Proposed Mining and Attribute Changes The truth is out there |

Hugh Ruka
Exploratio et Industria Morispatia
|
Posted - 2009.02.14 17:41:00 -
[47]
Originally by: BiggestT Slightly off topic but..
I find it hilarious that people claim that triangulation is useless in a 3d world (with depth), when thye use it themselves in a world of depth (the real world ofc).
Your using it right now!
Your eyes use triangulation to judge depth, distance and an overall location in a 3-D world.
As to why probes cannot use the same function I have no idea..seems very simple to implement...
omg ... fail
triangulation only works on a 2d plane ... your eyes are refocusing rapidly on different points on the object, and yous brain is just fooling around with the results. that's why you can perceive 3d space on a 2d surface when you are presented the right picture. if this worked as you say, stereograms would never work for humans ... --- SIG --- CSM: your support is needed ! |

Theqwert125
|
Posted - 2009.02.14 17:41:00 -
[48]
Edited by: Theqwert125 on 14/02/2009 17:41:45 Here is why you are wrong: the probes ONLY measure distance. This means that triANG(le)ulation will not work; you must use trilaterization, which, as stated in the article, results in two points, one above and one below the plane of the probes, requiring a fourth to eliminate the erroneous one. The ship does not count as a point of reference, as it does not have a probe's scanning equipment, and cannot detect things like wormholes. The dev, quite bluntly, misspoke when he said triangulation.
|

Sophia Truthspeaker
THE INTERNET. Goodfellas.
|
Posted - 2009.02.14 18:17:00 -
[49]
Originally by: Hugh Ruka omg ... fail
triangulation only works on a 2d plane ... your eyes are refocusing rapidly on different points on the object, and yous brain is just fooling around with the results. that's why you can perceive 3d space on a 2d surface when you are presented the right picture. if this worked as you say, stereograms would never work for humans ...
Yep, triangulation only works on a plane... pity that you can draw a plane through any three points you care to set... which means it works in 3d because it works in 2d ^^
_________ Proposed Mining and Attribute Changes The truth is out there |

Hugh Ruka
Exploratio et Industria Morispatia
|
Posted - 2009.02.14 18:23:00 -
[50]
Edited by: Hugh Ruka on 14/02/2009 18:23:57
Originally by: Sophia Truthspeaker
Originally by: Hugh Ruka omg ... fail
triangulation only works on a 2d plane ... your eyes are refocusing rapidly on different points on the object, and yous brain is just fooling around with the results. that's why you can perceive 3d space on a 2d surface when you are presented the right picture. if this worked as you say, stereograms would never work for humans ...
Yep, triangulation only works on a plane... pity that you can draw a plane through any three points you care to set... which means it works in 3d because it works in 2d ^^
it works in 2d but that does not mean it works in 3d for exact location. a plane is still not a point in space, it's a plane. and a plane contains an infinite number of points ... needle in haystack is freaking easy compare to that ... --- SIG --- CSM: your support is needed ! |
|

Jian Gi
Caldari Destructive Influence KenZoku
|
Posted - 2009.02.14 18:41:00 -
[51]
Actually this is one of the closest to RL aspects of EVE!
Ever wondered why you need at least 4 satellites to get a GPS fix ??
In principal measuring distance is fairly simple (and with a constant accuracy in regards to object ditance). On the other hand radial accuracy/discrimination is much harder.
|

Miss Moonwych
Formedian Shadows
|
Posted - 2009.02.14 18:48:00 -
[52]
Edited by: Miss Moonwych on 14/02/2009 18:53:52
Originally by: Kazang Plan and elevation view of the probes
A,B and C are the probes, the black lines and circles represent the scan result. Its a single distance. Which expands to a sphere. The red line in elevation is a radial result from 2 probes, it is a circle in plan. The 3rd probe's result will be a sphere tangential to the red cirle, giving a single result! The yellow and blue lines are the results of the 2 other probes with respect to each other.
Just as a reaction. Your error in thinking is in the above picture.
It should look like this (your left picture):
fixed pic
The black circles (actually two of them each) now represent the scan result on the plane-level of the target and the ghost target. The green dot represents the two possible locations for the target.
Basicly you were/are assuming that the target lies exactly in the A,B,C plane. Which most of the time wont be the case. Only when the target happpens to be perfectly on the same plane as the probes you need just 3 probes (already indicated by the devs). But when they're not on the same plane (99% of the case) you really need 4 probes (or warp to both if that were possible).
Hope that clears it up.
Regards,
M.M.
|

Sophia Truthspeaker
THE INTERNET. Goodfellas.
|
Posted - 2009.02.14 19:00:00 -
[53]
Originally by: Hugh Ruka it works in 2d but that does not mean it works in 3d for exact location. a plane is still not a point in space, it's a plane. and a plane contains an infinite number of points ... needle in haystack is freaking easy compare to that ...
Its pretty easy... imagine a table. On that table three marbles. You know the location of two marbles, and you know the direction you have to take to get from either one of the marbles to the third. Triangulation says you now know the location of the third marble.
To make it 3d, simply lift the table on one side. If the marbles down fall to the floor (because they are sticky or whatever) they will still form a triangle on the table. Let the side of the table fall down, but the marbles will hold their place, and now you got a triangle in 3d, showing you triangulation still works in 3d...
BTW, how do you think surveyors cartographed the land? ^^
_________ Proposed Mining and Attribute Changes The truth is out there |

Xianthar
STK Scientific The Initiative.
|
Posted - 2009.02.14 19:44:00 -
[54]
Assuming all fixed points are non co-located and reasonably spaced out around the unknown point
Distance to an unknown point from 1 known location (a probe) will place the unknown spot somewhere in a sphere around that probe with radius = the distance measured.
Distance to an unknown point from 2 known locations will place the unknown location somewhere on the circle than is the intersection of the spheres formed when looking at each probe separately
Distance to an unknown point from 3 known locations will place the unknown location as one of 2 points, the intersection of the spheres formed when looking at each probe separately
Distance to an unknown point from 4 known locations will place the unknown location at precisely 1 possible location.
Nothing is made up here, 4 locations are required to find a point in open 3d space.
Triangulation works on the surface of the earth to find "stuff" because with 3 known distances you get 2 possible locations for the unknown, if the known locations are properly spread out only 1 of those 2 possible points will be on the earths surface, thus you can rule out the other.
More known distances just help to account for inaccuracies in the distance measurement method which there is always some due to various issues. For instance the GPS system has to account for the fact that time travels slower on the earths surface than it does in orbit.
cheers
|

Hugh Ruka
Exploratio et Industria Morispatia
|
Posted - 2009.02.14 19:59:00 -
[55]
Originally by: Sophia Truthspeaker
Originally by: Hugh Ruka it works in 2d but that does not mean it works in 3d for exact location. a plane is still not a point in space, it's a plane. and a plane contains an infinite number of points ... needle in haystack is freaking easy compare to that ...
Its pretty easy... imagine a table. On that table three marbles. You know the location of two marbles, and you know the direction you have to take to get from either one of the marbles to the third. Triangulation says you now know the location of the third marble.
To make it 3d, simply lift the table on one side. If the marbles down fall to the floor (because they are sticky or whatever) they will still form a triangle on the table. Let the side of the table fall down, but the marbles will hold their place, and now you got a triangle in 3d, showing you triangulation still works in 3d...
BTW, how do you think surveyors cartographed the land? ^^
I give up ... --- SIG --- CSM: your support is needed ! |

Hugh Ruka
Exploratio et Industria Morispatia
|
Posted - 2009.02.14 20:02:00 -
[56]
Originally by: Xianthar
Triangulation works on the surface of the earth to find "stuff" because with 3 known distances you get 2 possible locations for the unknown, if the known locations are properly spread out only 1 of those 2 possible points will be on the earths surface, thus you can rule out the other.
More known distances just help to account for inaccuracies in the distance measurement method which there is always some due to various issues. For instance the GPS system has to account for the fact that time travels slower on the earths surface than it does in orbit.
cheers
triangulation on earth works only on land. since the height is given by the height of the surface, you only need to work 2 dimensions (a plane bent to form a sphere basicaly).
however it would not work underwater or in the air, as it would give an infinitely long line of possible locations :-) --- SIG --- CSM: your support is needed ! |

Raynardine
|
Posted - 2009.02.14 20:03:00 -
[57]
Originally by: Deva Blackfire
But it IS fun and quite simple. Problems are: 1. interface is slow and cumbersome. There are 10000 other ways you could move probes around which would be simpler and faster at the same time
I made a post on this, hasn't really gotten much attention beyond pageviews though... ------------ 1. Select a probe. 2. Press [m]ovement, or right click anywhere. 3. Hold shift+move the mouse up or down. 4. Right or left click to send the probe to that location.
This is an extremely intuitive method of moving objects in a 3D space. The only 'hard' part about it is figuring out which buttons do what. [ie - a new player doesn't know that holding shift will shift to the z axis.]
|

Hereon Herinnger
Gallente Nolra Corp
|
Posted - 2009.02.14 20:09:00 -
[58]
This isn't that complicated, people. You drop a bunch (number to follow) of probes. Each probe finds the DISTANCE from that probe to the anomaly. So with one probe you can narrow the location down to a sphere around the probe. Two intersecting spheres make a circle. Three resolve to two points (unless you're really really really lucky) and FOUR probes define a single point. Angles, magnitudes, and anything else have nothing to do with it. You only know the distance from the probe to the hit.
Now given the fact that the probes are not perfectly accurate, we must do this again with smaller radius probes. But I think we can all understand that part.
I think the balance is fine (except that some of the scan strengths should be increased so they are possible to pin down, at least with perfect skills and preferably without). |

Xianthar
STK Scientific The Initiative.
|
Posted - 2009.02.14 20:15:00 -
[59]
Originally by: Hugh Ruka
Originally by: Xianthar
Triangulation works on the surface of the earth to find "stuff" because with 3 known distances you get 2 possible locations for the unknown, if the known locations are properly spread out only 1 of those 2 possible points will be on the earths surface, thus you can rule out the other.
More known distances just help to account for inaccuracies in the distance measurement method which there is always some due to various issues. For instance the GPS system has to account for the fact that time travels slower on the earths surface than it does in orbit.
cheers
triangulation on earth works only on land. since the height is given by the height of the surface, you only need to work 2 dimensions (a plane bent to form a sphere basicaly).
however it would not work underwater or in the air, as it would give an infinitely long line of possible locations :-)
nope, you will not get an infinite line of points, you will get precisely 2 points in any reasonable arrangement, its simply the intersection of 3 spherical shells, which can never result in an infinite line.
cheers
|

small chimp
|
Posted - 2009.02.14 20:27:00 -
[60]
I think a dev stated somewhere that the probim interface is final. live with that!
|
|
|
|
|
Pages: 1 [2] 3 :: one page |
First page | Previous page | Next page | Last page |