Just wondering what is right:
My GPS Results filter settings in default mode, and in calculation
on cubic spline integration (10/20s Doppler)
shows my max ok, and 2 sec little smaller then max, what is right, but integration calculate only 10/20 sec average
When I switch to trapezoidal integration (2/10/20s Doppler)
there is 2 sec calculation, but my 2 sec is bigger then my max - not good
(but only in versions under 6, with latest 6.9 is like on KA72 and seems to works fine)
So I don't know will I leave on CUBIC or switch to TRAPEZOIDAL,
and someone told me that must be on TRAPEZ... like KA72, GPSAR and
GPSAR works - is it true or false ?
In GPS-Results the default filter setting is Cubic spline for 10 sec and 20 sec Doppler. The other results are calculated using the rectangular method (including the 2 secs). When I compare the two second for my last session in RealSpeed under Rectangular integration I get the same results as GPS-Results. The default setting for most (maybe all other) versions of RealSpeed is Trapezoidal, which usually give very, very close results to Cubic Spline for the 10 sec and 5 x 10 sec (commonly within a few thousandths of a knot).
Stick with the default settings in GPS-Results. They are the settings required for the GPS-SS records. 
More info on why trapezoidal looks wrong on this version of GPSResults is on this thread:
www.seabreeze.com.au/forums/Windsurfing/Gps/GPS-Results/
I compare my 5 session between gpsr 6.7 & KA72
Here is results:
test 1 - SESSION 1, Punat 10.09.:
* GPSR: 2sec 27.72, Average 24.53, Alpha 13.87, Milja 14.53
* KA72: 2sec 27.69, Average 24.52, Alpha 13.85, Milja 14.78
test 2 - SESSION 2, Punat 11.09.:
* GPSR: 2sec 26.09, Average 23.75, Alpha 15.09, Milja 15.31
* KA72: 2sec 26.01, Average 23.75, Alpha 15.09, Milja 15.31
test 3 - SESSION 3, Punat 07.10.:
* GPSR: 2sec 27.84, Average 24.82, Alpha 14.86, Milja 13.56
* KA72: 2sec 27.79, Average 24.81, Alpha 14.86, Milja 15.78
test 4 - SESSION 4, Punat 08.10.:
* GPSR: 2sec 28.28, Average 25.30, Alpha 14.93, Milja 15.75
* KA72: 2sec 28.26, Average 25.29, Alpha 14.92, Milja 15.75
test 5 - SESSION 5, Punat 16.09.:
* GPSR: 2sec 25.75, Average 24.13, Alpha 13.46, Milja 16.52
* KA72: 2sec 25.60, Average 24.12, Alpha 13.42, Milja 16.52
So, becouse I entered all my scores by hand from GPSR, seems that
I "cheated" in 0,0x in 2sec, and being loose some KT in nautical mile
Scores are so the same, as difference is minimal, that there is no problem at my end what to enter. So far, as I know, for GTC all softwares and ka72 is legal to enter
First of all, I will try to picture here that GPSR shows exact numbers as KA72 if is setted correctly.
GPSR is a professional program, and have integration to GPSS, maybe nex ver will have integration to GTC :), time will tell
Main difference and allways problem is score of 2sec. 2sec is not max speed, then speed you hold for 2 sec, counting from 0, 1, 2 - so you need 3 points to have correct 2sec. KA72 in that case working fine, but what bothering users in gpsr ?
Maybe my test will clear this:
As Nebb said, KA72 is the best way and the easiest to post score, and Nebb also said that filter in GPSR must be on trapezoidal setting to be the same as ka72 reding 2sec.
If you have older version of GPSR that will don't work, so check this out:
1. Get latest version of GPSR v6.90
2. If you have payed licence for version 6.xx, just install over old one, key will work, becouse in gpsr you pay for all newer version. There is no need to do nothing, just reinstall it. (If you payed for versio 5, contact Manfred for new key before doing upgrade - this work only if you payed for 6.xx key)
3. Now you have v 6.90
4. I open one file in default setting and it looks like that:
You see that gpsr is setted in default on cubic spline and 2sec is calculated
from 2 points (grey colour in the point box) - 2sec is 36,282
Now, try to switch cubic to trapezoidal.
In that case you will got this:
Now you see that 2sec is calculated from 3 points (3 grey colour line in point box)
2 sec in that sample is 36,13 KT
When I uploaded that same file on KA72, here is a result:
As you see, KA72 shows 36,13 as correct setted gpsr, no any difference in decimals. So, I trying to say that correct setted gpsr show the same numbers as ka72, and there is no theory that 2sec is bigger then on ka72.
Sorry 4 my english, I hope you understand what I trying to say here.
Update:
I also checked this same file with GPS Action Reply Pro, and in that case
2 sec is also 36,13 - so 3 times exact score, how more precise can be?
Nice work Warrior, this will help someone else for sure.
yes thanx
when i want to write so many words seems to my english block sometimes 
As Nebb said, KA72 is the best way and the easiest to post score
I actualy meens "the easiest and less painfull, quicker way, or something like that"
SO, IN THE END:
We all can use gpsr, gpsar, ka72 or realspeed to post ?
(as long as we know what parameters we read from this programs), right ?
And I wonder, will that my test worth something or not ?
Also check that same score with with gpsar and 2 sec was 36,13
3 time exactly same readings, guess set up is ok
Now is only the confusion: leave like that or switch to cubic and have little higher 2 sec ? (Asking that for other, not for myself), Score I tested with setup like that all show exact numbers in all 3 software I used
You get the same 2 sec speed in GPSAR Pro because it is also set to Trapezoidal calculation. Switch it to rectangular by un-ticking the box and see of you get the same result as cubic spline (rectangular for 2 sec) setting in GPS-Results. I am willing to bet it will be the same or within 0.01 knots anyhow.
Since in the past we haven't put much weight on the accuracy or validity of the 2 second score, we have not paid much attention to which method people use. It is really thought of as just for a bit of fun. However, we do now count this score in the GPSTC team rankings so maybe we should pay it a bit more attention??
I mean that when GPS-SS was set up, not much attention or credibility was given to the 2 sec figure. It was just a curiosity and does not figure in the main rankings. GPSTC changed that a bit because it is now counted as one of the rankings in it.
Currently, you post whichever you choose in GPSTC (Trapezoidal or Rectangular). I am saying, maybe we need to review that, perhaps for next years rankings?
ah, right
gpsss is more like based on 500met and similar
so, if trapezoidal settings in gpsr show exact score as ka72 and gpsar
i suggest that we who use gpsr set on trapezoidal
in that case ka72 post, gpsar post and gpsr post will have the same
2sec parameter, calculated on 3 points, and for correct 2sec you must have 3 points.
If users in gpsr have cubic, their 2sec is little bigger then on ka72 becouse cubic calculate 2sec from only 2 points
thanx sail ;)
I will also like to stay on default, on cubic spline
but what with test ? only in trapezoidal score is the same as ka72
with cubic is little higher ?
i don't want to anyone told me that I enter higher 2sec then ka72 calculate
I contacted Manfred (maker of GPSR) and he suggest to stay in default,
so in cubic spline In that mode, only 10s and 20c is calculated by cubic method and all other parameters are calculated by rectangular.
So, when is GPSR in default, every parameters is calculated in best method, so suggestion is leave it as is, no matter what my test proof or not
most important 10s category is computed by cubic spline by default, the other settings are just for comparison with other approaces. All other categories are computed by rectangular integration by default.
Trapezoidal integration gives systematic errors (too low speeds) for our typical peak-like maximum speeds.
I would use rectangular for the 2s category (this is the default in GPSResults).
(when is GPSR setted to cubic spline)
