I have mine on carpet, and at first had lots of problems keepingt the track straight. The track is sindgle lane which I put on plywood with sides to stop a car that might try to jump of the track. the wood part is actually wide enough to add a second late some day if I want. However, it has been up for a year now, and I found out a 2nd lane would be of no use to me. As I test, I want the running lane the same for each run.
What I found out quickly is that the carpet pyle makes the track creep. It creeps so slow that for a while you don't see it, for a couple of days even, especially if you don't have something that will allow you to check for creep.
Finally I put a tight string (like carpenters use) from one end to the other. And with that I could see a bit of creep on a daily basis.
I solved this by using a set of 1 by 2s attached to each section of the track and going all the way to the wall and firmly attached at each end.
IN my case, I had enough room to set the track parallel to the walls, about 4 feet away. I left the guide string up permanently so I can see if the supports need to be reset. I had to cut off the end piece of one track section in order to have a 4 foot stopping section (which I wish was two feet longer). Anyway my track is 40 feet long with another 4 foot stopping section.
When I send cars to race by proxy, I get a time just before shipping, and again just after I receive them back, then I create a ratio of the NPWDRL track to mine. Currently my ratio is .9925. Now I can get a good idea how new cars might perform at NPWDRL proxy events.
Maybe the creep depends upon what kind of carpet and how thick the pyle is. But be sure to start out using the guide string.
By the way, I use a DerbyMagic single lane timer on an aluminum track that I built. The front end of the timer is manual, but the back end is a beam. I have been using it for over two years now, first with a DerbyMagic plastic track, and later with the aluminum track that I built. With all the wood I used, the end cost was too high. With after the fact thinking, I should have just bought a BestTrack. However, in spite of my choosing the wrong direction cost-wise, My end result is perfect for me.