I just finished uploading files from recent Burrum and surprisingly one comes with excess of 35 knots 2S peak !?

Ok, I did change board to my speed board at late evening but that is still too good to be true.

File is available to view on KA72

www.ka72.com/Search/r/0/tf/2016-03-11/tt/2016-03-11

What was real my 2S speed then ?
BTW> There are two files: one till 5PM on IS97 and another after 5 on JP 56.

Hi Macro your original SBN file you posted on the 11th looks fine peak 26.650 (103min).
You posted a sbp file today which seems to be incomplete for the 11th (25min), im assuming the data has been overwritten by later sessions.
Have you posted from the same device.
either way i get the same peak with both files.
checked using gps results

Select to expand quote

vando said..
Hi Macro your original SBN file you posted on the 11th looks fine peak 26.650 (103min).
You posted a sbp file today which seems to be incomplete for the 11th (25min), im assuming the data has been overwritten by later sessions.
Have you posted from the same device.
either way i get the same peak with both files.
checked using gps results

Hi Mike, please have a look at recording after 5PM , when I did change a board.
Definitely speeds are higher then 26ktn.
My GPSResults comes with 31.4 knotts 2 second.

I did send message to Dylan to have a look at file...
There are two files for the day. I switched GPS off and on around 5 PM

High Macro, I get 31.325 with GPSResults, filters on doppler on.
Yet another reason why KA72 shouldn't be relied on!
Daffy had a 30+ NM but KA72 only gave him a 26

Select to expand quote

Macroscien said..

vando said..
Hi Macro your original SBN file you posted on the 11th looks fine peak 26.650 (103min).
You posted a sbp file today which seems to be incomplete for the 11th (25min), im assuming the data has been overwritten by later sessions.
Have you posted from the same device.
either way i get the same peak with both files.
checked using gps results

Hi Mike, please have a look at recording after 5PM , when I did change a board.
Definitely speeds are higher then 26ktn.
My GPSResults comes with 31.4 knotts 2 second.

I did send message to Dylan to have a look at file...
There are two files for the day. I switched GPS off and on around 5 PM

Yer but that was on the 12th not the 11th

In Gps Results make sure you untick the other days you don't want other wise it give the results for all the days.
Ka72 should give you a list of sessions then you just post the correct date.

Decrepit this is the session macro is inquiring about

gpsteamchallenge.com.au/sailor_session/show?date=2016-03-11&team=108

Select to expand quote

vando said..

Macroscien said..

vando said..
Hi Macro your original SBN file you posted on the 11th looks fine peak 26.650 (103min).
You posted a sbp file today which seems to be incomplete for the 11th (25min), im assuming the data has been overwritten by later sessions.
Have you posted from the same device.
either way i get the same peak with both files.
checked using gps results

Hi Mike, please have a look at recording after 5PM , when I did change a board.
Definitely speeds are higher then 26ktn.
My GPSResults comes with 31.4 knotts 2 second.

I did send message to Dylan to have a look at file...
There are two files for the day. I switched GPS off and on around 5 PM

Yer but that was on the 12th not the 11th

make sure you untick the other days you dont want other wise it give the results for all the days.

Yes Mike , you are right ! as usual It looks for me now that 31+ was recorded on next day , but somehow posted from KA to GPSTC on 11 not 12 March.
That is possibly my fault not KA as on my GT-31 i did performed save to card operation today , as I forget to put SD card.
All files and days then combined into one, at the beginning date ( i guess) .Anyway 35+ ktn still remaining mystery for me and I promise to you that my MacroBoat with 18 HP Tohatsu outboard is not able to falsify results so far 12 knotts was PB for my MacroBoat

That outboard is just there to fool us, no doubt got a inbuilt jet engine in there .
all good Macro

Select to expand quote

Macroscien said..

vando said..

Macroscien said..

vando said..
Hi Macro your original SBN file you posted on the 11th looks fine peak 26.650 (103min).
You posted a sbp file today which seems to be incomplete for the 11th (25min), im assuming the data has been overwritten by later sessions.
Have you posted from the same device.
either way i get the same peak with both files.
checked using gps results

Hi Mike, please have a look at recording after 5PM , when I did change a board.
Definitely speeds are higher then 26ktn.
My GPSResults comes with 31.4 knotts 2 second.

I did send message to Dylan to have a look at file...
There are two files for the day. I switched GPS off and on around 5 PM

Yer but that was on the 12th not the 11th

make sure you untick the other days you dont want other wise it give the results for all the days.

Yes Mike , you are right ! as usual It looks for me now that 31+ was recorded on next day , but somehow posted from KA to GPSTC on 11 not 12 March.
That is possibly my fault not KA as on my GT-31 i did performed save to card operation today , as I forget to put SD card.
All files and days then combined into one, at the beginning date ( i guess) .Anyway 35+ ktn still remaining mystery for me and I promise to you that my MacroBoat with 18 HP Tohatsu outboard is not able to falsify results

That outboard is just there to fool us, no doubt you got a inbuilt jet engine in there .
all good Macro

Vando I downloaded this file.
Macroscien 35.426 11mar 201610.0 Burrum Heads QLD, AU 2016-03-17 09 30

KA72 has these results.
2 Second Peak (kts): 35.427
5x10 Average (kts): 24.816
Top 5 5x10 speeds: (1) 26.341
Top 5 5x10 speeds: (2) 25.958
Top 5 5x10 speeds: (3) 25.249
Top 5 5x10 speeds: (4) 24.621
Top 5 5x10 speeds: (5) 21.913
Alpha 500 (kts): 16.099
Nautical Mile (kts): 23.801
100m peak (kts): 26.363
Total Distance (km): 15.978
5x10 Burrum 2016: 22.075
2 Sec Burrum 2016: 35.427
NM Burrum 2016: 23.801

My analysis of this file only gives that 31kt 2sec, I don't care whether it's from the 11th or the 17th, KA72 still gives 4kts more than I get.

Just checked the 2s for all the days on the file
11th I get 26.65
12th 31.325
15th 28.166
17th 25.951

Yes i agree 35 is incorrect.
But there is data missing from the sbp file on the 11th session time 25min
download the sbn he has there for the 11th there nothing wrong with that file session time 103min.
I think Macro has thrown a curve ball at KA72

Select to expand quote

vando said..
Yes i agree 35 is incorrect.
But there is data missing from the sbp file on the 11th session time 25min
download the sbn he has there for the 11th there nothing wrong with that file session time 103min.
I think Macro has thrown a curve ball at KA72

Yeeeh , right ... I have no other option now but back to Burrum and do this 35 knots properly
Definitely the site has the potential for the beginner like me to get results...
Give me a week or month and I will back with proper results...I have all gear to do so already
just need some wind and lucky

BTW I did observed some unexplained picks on downloaded from KA GPS files, but since not mine - never complained

I can see now what happen.
The file consist of 4 days of sailing.
Indeed for 11 March top speed was just 26.65 knots and 31.32 was on 12 March.
Where 35 knots comes from I have no idea as there is nothing during 4 days



It may suggest there is small glitch in KA72 algorithm.
It will be interesting if one day administrator will run bulk processing of all historical data to check if there were many similar glitches in the past.

yeeh it looks that problem is spread more often just short look on listing the same day 12 March and KA versus GPS Results brings more surprises ...








now everybody may check from time to time, if news are too good to be true

Nice detective work Macro. A very good example of the true spirit of the GPSTC

Well I set out this morning to try and figure where that 35kts came from. I know GPSResults ignores bad data, so I opened the file in GPSARPro. This has a handy feature of doing a sort in the speed column. Which of course I did, and it shows max speed over all those sessions is only 31.4something knots.
So then checked the trackpoint speed graph for signs of spikes, well there just aren't any.
Both doppler and trackpoint speed graphs max out at just over 31kts. So it looks like the 35kts has been conjured up out of thin air. This is a real puzzle, I can't figure it out.

In other KA files I've seen, bad data has been used to get an invalid result, well that's understandable it's just a different filter setting, (too low on KA72 in my opinion).
But there's no data I can see on macros file that goes over 31.5kts.

Yep, well done Macro, we may enlist your help sorting this stuff out.
Your definitely right, results need to be checked.

Anne forwarded an email to me today that Macro sent to the membership@windwanderers email address. I'm no longer on the WW committee, so best to use my regular email address (dylan@ka72.com) if you are trying to contact me, or better yet to post problem files into the specific problem files forum ( https://www.ka72.com/Forum/aff/10 ) that exists for that purpose. I have a handful of files there that have been brought to my attention. If there are any other odd ones out there (such as the aforementioned odd nm for Daffy) I hope that users will be proactive enough to post them there, rather than just sit back and snipe at the system.

A short time/distance error (such as one in a 2s peak) is much easier to diagnose and resolve than a long distance one such as a 1hr, so I tend to look at these first.

Macro raised a point which was:
> It will be interesting if one day administrator will run bulk processing of all historical data to check if there were many similar glitches in the past.

When a significant change is made to the KA72 engine, the version number of the engine is changed. Any time a track is subsequently viewed, the full result set is recalculated using the new version of the engine.

I have re-run the calculations on all files once before, when we moved to the current version of the site where historical results are cached. It was an extremely lengthy process which took several months in all, as there were over 100,000 track files in the database. Each file had to be retrieved from storage, reprocessed, and the results stored, without impacting too much on the normal usability of the site. The extra bandwidth and processing power required for the exercise cost me a few hundred $ too. So I'm hoping not to repeat the exercise too soon.

I always post my session to gpsteam challenge from ka72, but open them first in gpsresults, this has been going on for years, and I have never found any significant difference between the two result sets. Occasionally gps results cannot calculate the hour properly and comes up with an obviously wrong number (probably my filter settings), but ka72 always gets it.

Bugger Donald Trump, Dylan for president!!

I have on twice converted files like this that have shown wrong on ka72 and corrent in gps result to nmea format so the logged speeds can be read . KA72 makes up numbers from nothing its not a filtering error or anything like that there must be an error in how it converts the gt31 files to actual numbers.

So why the red thumbs, do you think Simon is making this up?????????????
Or is it that you can't stand the truth?

Select to expand quote

sillymilly said..
the truth is most of us are over you, daffy and co. always knocking KA72

We asked for a uniformed approach to posting but again you guys refuse to do that

sure you've added knowledge...but you and the other members of the know everything gang have created heaps of disdain

you sit back, accuse everyone of cheating, refuse to follow a uniformed approach, cast doubt and put shiz on ka72....all for what. ....
99% of us just wanna sail and are over you guys acting like police
Macro is a perfect example of what you have highlighted in nuumerous other threads....a guy posting crazy numbers that are just blatantly incorrect.....and you then turn around and blame ka72 rather than expect this guy to take some personal responsibility. ....

to blame ka72 for anything Macro has done is poor form
its clear to see Macro is a know everything/know nothing sorta man, so he fiddles around with his device and wouldnt really have a clue
eg he really thought he did the 35 peak...with his 24, 5 by 10
nuff said...clueless

all you guys have done in the last yr....do be dissmissive of ka72
accuse everyone whos not spending hrs drooling over there data of cheating
thrown real doubt over the integrity of the gpstc....

That is utter BS. Nobody *against* anything. All anybody is trying to do is to find the bugs in the hardware and software. The fact that 3 programs generate identical results, while the website generates something different.... should setoff your "whats going on here" alarm.

KA72 has been pm'ed on multiple occasions, asking for some bugs to be fixed. As is GPSResults, and the other tools. Whether any of those developers choose to fix the problems, is entirely up to them [ as it should be ].

By the way... calling someone clueless is actually offensive. If you were in the same room as the person, you wouldn't call them clueless.

Select to expand quote

sillymilly said..
to blame ka72 for anything Macro has done is poor form
its clear to see Macro is a know everything/know nothing sorta man, so he fiddles around with his device and wouldnt really have a clue
eg he really thought he did the 35 peak...with his 24, 5 by 10
nuff said...clueless

The first post and third post in this thread, was Macro specifically querying how such a peak.... so he didn't really think he did a 35kn peak.

Select to expand quote

sillymilly said..
We asked for a uniformed approach to posting but again you guys refuse to do that

Data has been posted here to show that some elements of both GPSResults and KA72, are wrong. If we use only a single instance of software to do the analysis, then these faults wont be identified. Surely you can see that?

Select to expand quote

sillymilly said..
the truth is most of us are over you, daffy and co. always knocking KA72

We asked for a uniformed approach to posting but again you guys refuse to do that

sure you've added knowledge...but you and the other members of the know everything gang have created heaps of disdain

you sit back, accuse everyone of cheating, refuse to follow a uniformed approach, cast doubt and put shiz on ka72....all for what. ....
99% of us just wanna sail and are over you guys acting like police
Macro is a perfect example of what you have highlighted in nuumerous other threads....a guy posting crazy numbers that are just blatantly incorrect.....and you then turn around and blame ka72 rather than expect this guy to take some personal responsibility. ....

to blame ka72 for anything Macro has done is poor form
its clear to see Macro is a know everything/know nothing sorta man, so he fiddles around with his device and wouldnt really have a clue
eg he really thought he did the 35 peak...with his 24, 5 by 10
nuff said...clueless

all you guys have done in the last yr....do be dissmissive of ka72
accuse everyone whos not spending hrs drooling over there data of cheating
thrown real doubt over the integrity of the gpstc....

For ages i was all for the compulsory ka72 concept I had wondered why some of the guys into the technical side of the gps never seemed that keen ,my logic was that even if it wasnt 100% atleast it would be even for every one. But then i realised why they have that attitude ,from what i can tell it actually doesnt work all the time and its not a question of filters not catching errors in the gps or how alphas are interpreted but it creating numbers that never existed.To find this i used gps babel with all filters turned off to convert the files to nmea then sorted the speeds in excel ( maybe my process is wrong ?) also both times i have seen this happen the gps didnt even display the peak speed shown in ka72.
I dont think any one likes to bring this up because we all apreciate all the work that has gone into this free service so no one want to talk negativly about it but people should be aware that some times it does get it wrong. I have only seen it twice and only with gt31 files but it may happen with others as well. I normally upload with ka72 and dont look at it in gps results unless it was a fast day.Im sure these problems will be fixed in time sometimes these sort of things are hard to work out.

I just noticed an interesting point " silly milly " brand new account your not the same super tough guy that was going to beat me up and put photos on the internet are you ? multiple account internet warrior ? this is probably the case so every one ignor my red thumbs there probably all this one guy with different accounts.

Select to expand quote

mathew said..

sillymilly said..
to blame ka72 for anything Macro has done is poor form
its clear to see Macro is a know everything/know nothing sorta man, so he fiddles around with his device and wouldnt really have a clue
eg he really thought he did the 35 peak...with his 24, 5 by 10
nuff said...clueless

The first post and third post in this thread, was Macro specifically querying how such a peak.... so he didn't really think he did a 35kn peak.

Thank you Matthew
I hope I didn't offend anybody rising the topic on our GPS discrepancy readings.But as smilly suggest for sailors like me, to disregard KA and post data directly, overwrite and correct what KA comes with - will be like a cheating or expression that I know better. But I don't . I am sure that small glitch will be resolved soon and that also could be useful outcome from my post.
If I just correct data by myself , nobody will never know ( outside this expert group). As for smilly to don't have even pleasure to sail on the same water , not just seating in same room. Hopefully we will never meet as Australia is so big....

BTW For example the data from SD card I posted from Burrum Heads directly from Android tablet- so any validation with GPS Results are outside the question is such occasion. One may use common sense to filter off 35 knotts, but that is main reason to get into places like Burrum to have this 35+

Then one clue that may help resolve that problem could be that this specifically error was created when submitted file was download from GPS with save to card command. One big file containing four days of sailing. Not just everyday files send directly from SD Card.Maybe GPS unit somehow remember all Max speeds ( even from driving car around) and such occasion dump on this command to resultant file ?

Select to expand quote

sillymilly said..

Macro is a perfect example of what you have highlighted in nuumerous other threads....a guy posting crazy numbers that are just blatantly incorrect...

If you provide us with you nick name for KA and GPSTC I could do for you free service to check how many knots you add to your posting...
by mistake or even most likely on purpose

all in all we can only hope Dylan can find time to fix thoses glitches, (I had an awesome 115 knts that I corrected while posting to gpstc ).

No point in getting aggressive, a little respect towards all the free work from Dylan isn't much to ask ?

Select to expand quote

sillymilly said..

its clear to see Macro is a know everything/know nothing sorta man

I am not necessary disagree with this statement.
This is exactly what the greatest philosophers used to say about them self.
But that doesn't simply imply that SmellyWilly known suddenly anything more or better.
Most likely the opposite is true.

As to KA I am the greatest admirer of Dylan hard work and magnificent result as KA72 site with all this unique futures and capability .

We are here SmellyWilly, to help Dylan resolve the problem and soon errors like that remain remote past.

On humorous side I have my own explanation to what happen here.

It is magic.




Since I consider myself the greatest supporter of AI on whole planet,
those poor babies of mine, smart computers networks around the world wished somehow reward me for moral support.

Since I am not the greatest sailor they did try and found the way to improve my standings and crank computations on supercomputers powering KA.

Thank you my babies but I don't need this sort of support from you. I will crack this 35 knots anyway one day

Bye bye Milsy

Just posting to let everyone know that I've had a look at Macro's file and worked out what the issue is.

There is a bug that was introduced when optimisations were created for GW52s. The effect of the bug is that the very first 2s peak of a file is calculated as if it was over two seconds, not three seconds. This artificially inflates the very first 2s value in the file by a factor of 3/2.

Since Macro was sailing quite fast when the file started, it's turned what should have been 23.6kts into 35.4kts.

Normally at the beginning of a file you are not sailing fast. It's only because the track was a truncated SBP file, and Macro was also sailing fast at that time, that the issue was noticable.

Thanks for pointing it out. I've updated the software, will check some other files that are also experiencing issues, and report back when it's updated on the site.

interesting point Dylan, I suspect I could be in the same boat (not solar powered) as Macro as I sometimes start recording on the go (cause I forgot to start the device on the beach)