The difference between "running" and "training" has for me been defined by whether I visit the track on a weekly basis. Arriving at the track provides a mental shift that says "you are here to run." Minimizing variables, it is a good fitness gauge. And the track facilitates steady improvement in things like speed and form.
I've been away from the track for a long, long time. Starting three weeks ago I've been going back
Track work starts with a two-mile time trial, run as if it were a race. Two miles is short enough to minimize variables but long enough to gauge fitness. (A mature runner can find a flat 5K race and accomplish the same goal.) Ideally one would perform one of these trials every spring and every fall.
The two-mile time trial provides a marker for the absolute level of fitness, reflected by the VDOT number. That absolute fitness level then tells one what paces to run for various training distances and helps to predict the pace one might expect to run in an upcoming race.
Here is the VDOT calculator. You enter your distance and your time and press "compute." My 12:20 2-mile TT gives me a 50.6 VDOT. Now I consult Jack Daniels' (not that Jack Daniels -- this one!) running pace charts. (The preceding link is to a Runners' World article reporting more or less the same thing I am reporting here and drawn from Jack Daniels' book and many free internet resources.) I learn that at a 50.6 VDOT I am at the fitness level that might produce a 20' 5K, a 40:30 10K, and a 3:08 marathon. (Some interpolation required to produce those numbers.) None of those surprise me based on recent years' experience.
From the 50.6 VDOT I also learn my "tempo," "interval," and "easy" paces. In his book Jack Daniels offers a tad more granularity than does the RW article. Again interpolating, my "tempo" pace is somewhere around 6:44/mile. At the track each week I run distances of 800m, 1 kilometer, 1 mile, 2 miles, up to 20' steady. All distances are run at tempo pace. (Different distances to keep things interesting.) 1' rest between intervals. The biggest mental hurdle to these workouts is that they are not that hard. I am confident I could run three or four miles well faster than a 6:44 pace with 1' rest -- I just ran 2 at a 6:10 pace, after all -- but the common and naive mistake is to see the track as a place to prove oneself instead of to improve oneself. (Just made up that catchy coach's slogan.)
At some point the assignment at the track changes and I will move to running shorter and faster with more rest. That is where "interval pace" comes in. I'm not there yet.
With regard to race-time predictions, perhaps needless to say, predictions based on the VDOT does not mean one can run any of those races tomorrow. It does mean that with training for the particular distance those are the times one might anticipate.
Of course, it is not that simple. Because with training one not just acquires the ability to cover the distance but one also improves absolute fitness level. Quoting a commenter on the Runners' World article linked above,
"Nice &simple explanation. but after 6 weeks shall we again go for
test and find new vdot and work on those numbers every time?"
Bingo. Maybe the next test is actually a (flat and not windy) 5K or 8K, because one will have the training to make those distances meaningful predictors. This will probably happen organically if one includes a series of B races built into a training program for the big fall marathon.
Another way of looking at this is to say the VDOT does not predict a future race at all. It predicts a current race. For a future race one needs to understand the pace of VDOT improvement, which presumably starts fast and levels off as one approaches maximum possible fitness. (I assume none of us at RP are approaching our biological capacity, so with steady training our improvement will probably be pretty constant.) A Runners' World forum relates observed improvements between 0.7/6 weeks and 1.5/6 weeks. Taking the more conservative of those and starting from my 50.6 VDOT reported above, I might hope to realize a VDOT of 52 while training for a September 1/2 marathon, for a predicted finish time in the 1:28 range. That sounds about right -- and not too bad of a goal!