GPS ELEVATION AND DISTANCE CALCULATION

Long ago, trying to decide in which site to upload my tracks, I accidentally made the following discovery. Any site or software,for the same gpx file, gave quite different results, especially regarding the altitude and the total distance.The scientist in me woke up and began to search. I contacted the owners of various sites (Garmin, everytrail) etc, but everyone was telling his own. Generally the answers did not satisfy. I searched the internet and found that there are others with the same questions.The explanation, after some searching, it is ultimately very simple and is as follows. The total elevation difference emerges by adding the individual ascents. This is done by using an algorithm, which determines the ascents and aggregates. The problem arises because each software uses its own algorithm. Now you say and which is the right one? The answer is none and all.The gps track record and store the position of each point. All points together makes a track. These points also include noise. That is why they have created different algorithms, which try to eliminate this noise, and also the fault of the gps. Essentially they are smoothing the elevation of our track, or else they are correcting it. If left uncorrected is definitely wrong. At this point to say that either use a barometric altimeter gps or not, the result is the same. Again uploading our files in different software, would get different results.What I ultimately did is that. I uploaded some of my tracks at known sites, software and compared them together.The result is tragic. The differences are between 50 and 1000 meters in elevation. What I noticed is that the only showing similarities between them (I would say that is exactly the same) are the Garmin conect using elevation correction and sporttracks. They also gives the correct distances, the ona that it is showing on the gps screen.One of my experiments was this. Compared the elevation of the route of the MTB race in Xirolivado, with the one of my friend Geron, which has a barometric altimeter gps. The Garmin conect and sporttrakcs were the only ones that gave identical results. (My 60cx doesn't have barometric altimeter).Another experiment was to record the same track twice on different days and compare the gpx files between the. Also Garmin conect and sporttracks gave the same numbers. All other out of place and time. So I came to the conclusion that these two are the most reliable. And do not forget that Mom Garmin support and endorse the sporttracks reliability. They surely know better. They are not accidentally, t he dominant manufacturer of gps in the world.Now on the distances. The gps calculate the distance in two dimensions rather than three. This means that when an uphill climb, the device calculates the horizontal distance and not the one we did. So we have error. The question is how large this error is. Suppose an uphill climb and the gps says it is 5 km distance and 300 meters of elevation. According to the Pythagorean theorem (the square of the hypotenuse equals the sum of the squares of the other two sides) the actual distance traveled is 5009 meters. Practically negligible for ordinary mortals. And a more extreme example, even the uphill we did is 100 kilometers and the elevation difference 3. The actual distance traveled is 100 045 meters. An error of 45 meters to 100 kilometers. I do not think are worth mentioning. Surely the bigger the slope and reduced the distance increases the error, eg a climber may see crazy numbers.And do not forget that top athletes of cycling and running are using gps in training and races.