@mathteacher thank you for posting for sure. I will enjoy reading this tonight. I did see it was 2011 but it should help us all see where we need to be focused. I think from 2011 to now they made a few changes wording mostly as if my memory serves me correct it was almost impossible not completely impossible to meet the standards of precision in some situations where small boundaries lot surveys and or property corners were very close together on the boundary. It is not uncommon to have a good size boundary here in the east say 15 acres more or less but on that boundary to have property corners only a few feet from one another. In that case or situation it usually blows up the rpp between those sometimes of allowable. Just to small a distance. There is a clause in the new standards that gives a surveyor an out in these situations because of terrain and such size shape all plays into this.
I still want to know the the proper procedure is for RPP, but I also don't understand why tight RPP matters for boundary/title lines in 2023 vs early 1800's techniques?
?ÿ
Because it's 2023 (almost) and not the 1800s any more? The monuments are the monuments no matter when they are measured, or how tight. I don't see a problem with having a reasonable standard for the measurement part of our profession, based upon the tools of our era.
2cm + 50ppm is pretty reasonable, and isn't very tight at all with modern equipment. Contractors with generic hardware store equipment can "measure better" than than the 1800s survey-specific equipment.
I would be pretty annoyed if I wanted to build on my property and the surveyor I hired for staking said "Oh we can measure to less than two centimeters relative precision with 95% confidence, but since the original surveyor was only able to get within two feet, that's all I'm going to give you for property lines."
To put it less flippantly, technical advancements don't automatically override the professional work we do. Most professionals use industry standard tools without compromising their professional duties - no reason we cannot do the same. Having some basic technical standards that we all agree are attainable and provable is a good thing.
?ÿ
RMS (what TBC does) is the way to compute RPP between points. We are combining error quantities of two points - the semi-major axes of the relative error ellipses - which per every statistics text means we add the squares of the values then take the square root of the whole thing.
which per every statistics text means we add the squares of the values then take the square root of the whole thing.
I might just be out to lunch here. Shouldn't RMS be the root of the mean square? Should be dividing the sum of squares by n before taking the root if that's the case?
I might just be out to lunch here. Shouldn't RMS be the root of the mean square? Should be dividing the sum of squares by n before taking the root if that's the case?
If you're out to lunch, then we must be at the same table.?ÿ
?ÿ
RMS:
?ÿ
From Trimble Help Files:
Relative positional precision is defined as the RMS value of the semi-major axis values of two points for which an error ellipse has been computed. Allowable precision is computed from the allowable tolerance and horizontal distance between the points. The ratio of these form the precision ratio which must be 1.0 or less to pass.
?ÿ
From the Ghilani paper:
Note that the standards do not state how the allowable precision is to be calculated.
Using standard statistical procedures, the proper computation of the maximum allowable semimajor axis for the 95% error ellipse is based on the sum of the squares of the two error components as shown in Equation (1).
So if this is how you calculate allowable rpp, then this is also how you calculate your test rpp.
Also, the Ghilani paper NEVER goes on to use any aspect of the error ellipse for the 2nd point being tested against. (Have we been looking at this all wrong?)
So in our case:
Allowable RPP = sqrt((0.07')^2 + (0.005')^2) = 0.07017'.
Point 1 RPP (with respect to point 2) = sqrt((0.02')^2 + (0.005')^2) = 0.021' (passes)
Point 2 RPP (with respect to point 1) = sqrt((0.03')^2 + (0.005')^2) = 0.030' (passes)
It's a bit odd to see at first, but depending on which way you measure the line, you will get a different RPP. But that makes sense given that the the concept is relative to the starting point with a different error ellipse.
It appears that Trimble's RPP definition (RMS of 2 points) does not align with Ghilani RPP definition (sum of the squares of fixed error and scalar error), and there is still quite a disparity across the board about how to compute rpp.
?ÿ
"Oh we can measure to less than two centimeters relative precision with 95% confidence, but since the original surveyor was only able to get within two feet, that's all I'm going to give you for property lines."
I understand and agree with much of your sentiment from your last post to me, but you stepped right into the "accuracy vs. precision" trap.
Which I maintain is one of the key reasons they switched from RPA to RPP.
Whose truth are (were?) we staking to, and how do you prove it?
Surveying: 75% "art", 25% "science".
@michigan-left nice. Looks as though Trimble based on who we know is slightly different than Dr. G. Paper. Also possible different than other software packages. ?ÿI will read this paper and get my findings an pass it off at the next chapter meeting to our nsps representative and ask him to give it to them so they can possibly add this or what definition they want us to use. ?ÿI know me and him sat down at dinner once my rep. And asked him is what i was doing correct and what tbc was doing. He said you are probably doimore than most surveyors. He said most are not even doing a least squares if they pass the precision closure by state standards that??s probably it 1:20000 etc. ?ÿif I ever get caught up on farm chores i will have to look up that video where someone represents the alta standards was describing it. It is a work in progress. They added new language to the table A and such lately. To help better protect the surveyor. I do agree with your statement 75% art 25% science. My first survey course in college the very first definition was written on chalk board. Surveying is an art and science????.. ?ÿand he kept that little phrase up on chalkboard for every class . ?ÿIt was exam and quize question on almost every test in some form. But that was in early 90??s. He taught us to use a scale and then taught us how to rough ck by sketching any units with our finger joints as an rough scale. ?ÿShowed us how we could often check our math just by drawing a little sketch to a scale to see if it made sense. It has saved me many times. Its not accurate but would show a major blunder sometimes. He said as a surveyor you need to learn to measure with body parts sometimes. He said because that finger your pacing your arm and foot you will not leave at the office. The transit tape and other items you will leave or forget or break or drop. ?ÿI think he helped moses part the nile. But was an awesome teacher. ?ÿMade you love what he taught.
I understand and agree with much of your sentiment from your last post to me, but you stepped right into the "accuracy vs. precision" trap.
How so? Older tech was/is still relative positioning, boundaries were/are still measured relative to themselves rather than to a national network/datum or absolute values.
ALTA specs specifically mention a minimally constrained LSA, which is by definition relative (internal) accuracy. True accuracy is technically unknowable, which is why we instead report results with reference to a standard - in this case, 2cm + 50ppm.
With respect to ALTAs, accuracy is relative (within the context of the current survey only), and it is a pass/fail metric. Did we meet 2cm + 50 ppm or not?
Whose truth are (were?) we staking to, and how do you prove it?
For me, the "truth" is the monuments - is their pedigree such that they may be accepted as marking the corner position? - and the "proof" is the results of our minimally constrained LSA, demonstrating that the held monuments were observed at or better than a specified technical standard.
The truth and proof together are what comprise a professional opinion. One surveyor's opinion (whether done in 1800 or 2023 with very different techniques) doesn't necessarily outweigh the other's. They might have distances, bearings and areas that are different, and perhaps different monuments shown or held, but describe the exact same parcel of land.
There may very well be material discrepancies between surveys, but ALTA standards only require that we document how our opinion differs from the record.
?ÿ
?ÿ
Also, I was wrong, RMSE is not correct. We are evaluating the error of a line, which is defined by two points, specifically the difference between them by subtracting one value from the other.
So then combining the error of those points falls under the category of computing the error of a sum (difference between two different points), not multiple observed quantities of the same point (which would then require RMSE). This is why using the square root of the sum of the squared values is correct from a statistics standpoint.
?ÿ
(That's all my opinion anyways, not trying to say anyone else is necessarily wrong, but the above is how I interpret the ALTA specs and how they fit in with the duties of a PLS performing an ALTA survey.)
@rover83 hey you and others opinions are what helps us all learn. I always enjoy your writings and your perspective. ?ÿThats how we all learn and get better by asking questions and challenging each other. ?ÿI do wish I could sit down with you for a few days on tbc. For sure. Every thing you have helped me or other with has increased my ability to get better at that software for sure. I wish I could find my books from control surveys though. It was an old but good one with lots of references to many authors. It was a dod book so had stuff from ngs well name was us coast and geodetic surveys then but other agencies and authors on things. Old paper back type. It was my go to for every control network design. It was more of a guide than book I guess but gave examples and then documentation on why and what could go wrong etc. so when tbc runs the alta report after reading dr g paper do you think Trimble has it wrong or just uses a different approach and is sufficient.
Of you start from a point and traverse in opposite directions to two other end points, then sqrt sum of squares of those points' semi major axes is a good estimate of the likely error in the computed length of the line between them.
If you make a measurement between those two points, the best estimate of the error in the length is much smaller, mostly determined by the centering and EDM errors of that length measurement, even though you don't know their coordinates that accurately.
That's why StarNet reports both coordinate and length error estimates.
@bill93 oh so what was shown by starnet previously was coords error maybe thats why we were a little different in computation. The paper that mathteacher posted and formula from that paper by Dr. G seems to be slightly different from what i see TBC doing its very close maybe starnet has a little different as well. For the allowable and what is. Rover posted the allowable and it seems taking the square root at the end for allowable is one part tbc and i were missing. I was doing everything except taking the square root. On allowable. ?ÿI am sure if i dig hard enough we could find different authors on how to do it . ?ÿI think from the post though and dr g is nsps does not define it is the key. They should maybe define exactly what they intend so we and manufacturers are all on same page.
@bill93?ÿ
That's why StarNet reports both coordinate and length error estimates.
Bingo.
Not sure why there isn't more discussion on the alternative (from ALTA):
Alternatively, Relative Positional Precision can be estimated by the standard deviation of the distance between the monument or witness marking any boundary corner of the surveyed property and the monument or witness marking an immediately adjacent boundary corner of the surveyed property (called local accuracy) that can be computed using the full covariance matrix of the coordinate inverse between any given pair of points, understanding that Relative Positional Precision is based on the 95 percent confidence level, or approximately 2 standard deviations.
You can use an inline option in Starnet to force the software to compute Tolerance Checks between specific pairs,
Here is my input
#MD MINIMUM STANDARD TOLERANCE CHECK 0.07'+50PPM
.PTOL 502LOC-314 314-512SO 512SO-522SO 522SO-523SO 523SO-300
.PTOL 300-501LOC 501LOC-500LOC 501LOC-502LOC
I pull the stakeout data into startnet and reran the adjustment with this inline option to get a report alone the boundary lines as staked.
Works nicely.