May be just calculate SD for each of 3 repeated measurements, then for all of 'SDs of 3' calculate some stats and use it? But from theoretical point of view problem of least squares adjustment just can contain all the data, i.e. all repeated measurements while distributing errors, and as secondary result of adjustment will provide the SD. But it is needed to modify loop-closure code, ofcourse, probably using 'ceres-solver lib'... Just for reference of theoretical thinking above, here is a nice book https://www.amazon.com/Observations-least-squares-civil-engineering/dp/07002... On Sun, Aug 5, 2018 at 12:16 PM, Wookey via Therion <therion@speleo.sk> wrote:

On 2018-08-05 08:44 +0200, Graham Mullan wrote:

...

I am not sufficiently versed in the maths to suggest what figures might be used, but I do want to know why you think that the 3-times feature might affect the expected error for the angles but not for the length?

The 3-times thing greatly reduces blunders. Blunders are usually in compass and clino readings, rather than length readings. But you are right that the 3-times measurement improves all 3 readings (I assume that's what you were implying).

Now the SD doesn't represent blunder frequency as it's really about expected mesurement error, but in practice we use it for blunder distribution too, so that's why a lower SD for surveys done this way makes sense to me. If we put in all three versions of the leg then we could leave the SDs the same, but that's not often done (depending how you get your data out of a distoX).

Similar considerations apply to backsighted surveys, where both sets of readings _are_ normally entered.

Wookey -- Principal hats: Linaro, Debian, Wookware, ARM http://wookware.org/

_______________________________________________ Therion mailing list Therion@speleo.sk https://mailman.speleo.sk/listinfo/therion

I try to get people to rotate the distox when taking some/all readings depending on there keenness and how easy it is*. The distox records and sends its rotation to pockettopo and I put this into the therion/survex file. So should the rotated leg have a different sd to the none rotated, probably. It would be useful if therion/survex could do this automatically from the readings but I guess that would require there to be a rotation field in *data *rotation is a partial check on calibration and generally picks up local magnetic interference, so at a minimum rotate every 5 or whenever we suspect metal. Andrew On Wed, 15 Aug 2018, 15:23 Footleg, <drfootleg@gmail.com> wrote:

In my own experience, the most common blunder which the 'measure 3 times' method with DistoX2 catches is mis-read distance. The Disto tends to be used in the same orientation for all 3 shots, so the compass and clino tend to be well aligned in all 3 (regardless of how well calibrated the DistoX is, or is there are local magnetic fields throwing the compass). But the laser does sometimes miss the target and overshoot, resulting in one of the 3 shots being out of tolerance for PocketTopo to count them as a valid leg. Of course catching blunders so they never make it into the data (we would reshoot when it happens) is not the same as the SD in the measured data.

The way we catch bad calibration problems or local magnetic interference is to take backshots (3 of them). This has saved serious errors on multiple occasions which would not have been detected taking shots in one direction alone.

Footleg -- Survex https://lists.survex.com/mailman/listinfo/survex