As an engineer, there's more to this than just the raw numbers being "accurate".
It's like the difference between "accuracy" and "precision" when shooting. In shooting, "accuracy" is how close you are to the intended target (like the bullseye on the target). "Precision", however, is how close multiple impacts are to each other.
For a chronograph, there's "accuracy", which is how close to the actual velocity of a projectile the readouts are; and then there's "repeatability", which is the range of variation in results when performing the same action. I'm sure there are statistical terms which could be applied here...like "mean", "average", "absolute", "maximum", "minimum", etc.
If you're really interested in accuracy, then you'd have to have your instrument calibrated against some standard. And even then, other local factors could influence this.
I'm not a reloader (yet), but as an engineer I can point out that "repeatability" will give you very meaningful information in your reloading efforts.
What you are ultimately looking for when reloading is producing a load which gives you reliable, repeatable, and acceptable terminal ballistics. The load cycles in your gun with no issues, it doesn't tumble, it hits the target down range with an acceptable accuracy and precision, the round expands adequately (if hollowpoint), etc.
For any given loading (bullet, powder, primer, cartridge, firearm combination) which produces these acceptable characteristics, there will be some spread of bullet velocities associated with it. So long as you use the same chronograph, those numbers will mean something to you.
So, if you use bullet X, powder Y at Z amount, and so forth, everything you load this way should produce a fairly consistent velocity.
If you want to increase the velocity for more performance, then you change something...like the powder or the powder loading. You do this incrementally, while observing the chronograph results and the terminal ballistics. If you get to a point where the bullet holes in the target aren't looking right, then you know that exceeding that particular velocity AS SHOWN ON YOUR CHRONOGRAPH is a limit for that combination of loading factors.
This will serve you quite well as a reloader.
However, if you should manage to shoot your chronograph one day and need to buy a new one (which I understand happens now and again), then your new chronograph might produce readings a little higher or lower on average than your old one. Whether this is enough to make a significant difference depends on how much they differ.
If you get a new chronograph, then you'd need to do some comparisons and see what the difference is. Regardless, that difference would only be used as a guide in your reloading anyway, because as you go forward you'll be developing loads based on your new chronograph's data just as you did with your previous one.
The only way around this is having your chronographs calibrated against some standard, just like for torque wrenches, multimeters, calipers, micrometers, etc. THEN you can lay claim to a known accuracy and compare values from one chronograph to another.