As with anything else, the more welding I do, the better my welds are becoming. It also means I'm figuring out those little tricks that create significant improvement in various specific welding scenarios that you don't even think about until after throwing yourself into those scenarios. By the time I'm done with the chassis, I might actually make a half way decent welder! Of course in the mean time, this means I get to look back on all of the welds up to the point of each new revelation and try to force myself not to simply start over. It may however mean that the earlier sections of chassis will be getting a little additional gusseting...Just in case.
On the other hand, my overall results have been good with keeping weld distortion to a minimum so far. After tacking the section all together, I pull it out of the 'jig' so that I can flip and rotate it around into whatever position will allow me to reach any given joint the most easily. My technique starts off with welding all of the fillet welds on the inside corners for the main boxed segments first, alternating corners around the chassis. Then doing the same with the outside corners. After that is the fillet welds on all of the diagonals. And last comes all of the the butt welds on either side. Really the primary source of what distortion I have seen appears to be coming from any minor variances in tube fitment. Wedge a tube in too tight, and when welded it bows the tubes on either side out. Leave a little gap, and the weld will pull it closed. None of this shows up well in pictures though.
However, I do have a picture illustrating one other method I'm using. I find the hardest to reach weld for every tube joint, and figure out what position both of us (the chassis and myself) need to be in to most easily weld it. This includes determining which direction the weld will go. I then mark that next to the joint. Every other joint around that tube intersection is then welded moving the same direction, such that the coldest part of the weld at the beginning overlaps the hottest part at the end of a previous weld. Alternatively the hottest part of the weld at the end overlaps the coldest part at the beginning a previous weld.