What makes a good test track?

Greetings to All!
I'm essentially a complete novice to 1/64 racing and modding (I dabbled in the game Gaslands a while back), and reading the article on timers last night got me thinking a lot about doing a test track. While I'm planning out my first race track already, the need for some type of test track for my mods also became clear to me since I can see that using my own track for testing would only optimize my mods for my own track. Still, I have no real idea about what would work and what wouldn't, what to include, dimensions, etc.
Since there are so many knowlegeable and successful folks on here, I thought I'd throw out the question "What makes a good test track?"
Any and all comments are appreciated.
Happy Racing!
-UncleBuster
Discussion
For me, i've tested with a tempory track of either crash racers or hotwheels orange across the livingroom and went through all my cars to find my fastest stock car( a hotwheels evo) then i base my builds on whether they can beat that evo 2/3. i am in the process of building 2 tracks in fixed locations in the garden, constant slope drag for pure speed testing and i hope to build a diorama track too which whilst not a true test will tell me how to expect things should do at other crash racers tracks , and whi;st every track is unique the charactaristics are tellable , but as i get deeper i find i'm favouring some places over others
I just have a two-lane I set up. I like the start gate from www.target.com/p/hot-wheels-flying-customs-race-and-jump-trackset/-/A-79391425, and I like the 2-foot track you can find a Dollar Store as they give less connections to snag on. Without a timer you can to A-B testing to find the fastest, but a timer is pretty essential if you want to figure out what mods make you faster.
Building your own timer is not exactly easy, but if you are comfortable with soldering and not afraid to mess with Arduino, I've written a guide on how infa-red gates work and can be built. It is geared to Nerf, but applies to any small, fast moving toys: www.instructables.com/Arduino-for-Nerf-Ballistic-Chronograph/
My test "track" is a one-foot section of my desk in front of my keyboard. If the car rolls straight, smooth, and quiet, then I got something decent.
If you're talking dragstrip racing then a good test track would be the same thing. I'd say at minimum about 10-feet length. If you're not that pressed for space then go as long as you can go. Build what you want and it'll suffice for whatever testing you want to do. Check out the Track Directory if you want to see a good spread of track shapes and sizes and get some ideas.
Here's the thing though, is that no matter what you build, it will only give you a hint at how your car will perform on another track. I mean, physics is physics and if your car rolls smooth and is the fastest car in your fleet, then it'll be fast on another track.
You might not need a separate "test track" at all. The drag strip you make for your regular racing will be just as good as anything you make specifically for testing.
Since there is no standard for a track specs, what runs good on yours may run (very) different on another. A steeper drop or longer track or shorter drop will all change your car but in the end, a fast car is a fast car (in a straight line).
Now, with open tracks that twist and turn, there's no good way to test your car for those at all. There are some general rules to follow with weight placement in your car and whatnot, but overall, it's really just luck on those tracks - they're all so custom and unique. I wouldn't bother trying to test for open track racing. Just because your car is fast on a drag strip doesn't mean it'll do well on an open track at all (might be the opposite).