Are "prediction workouts" useful?

For prediction workouts taken from a set training plan or another coach they may not line up as well with your specific athlete. However, if they come from a reputable source it should be a good point of reference for the race itself. Also, it can be good to start something like this in your athlete's training program, see how the race goes, and then incorporate it in future seasons so that the prediction workout correlates with the race and can be done in similar times down the road, then compare feeling and have a better idea of race fitness.

Ultimately a race simulating workout is just meant to practice feeling and give this kind of reference point. I have always liked workouts like 8-10 x 400 off 1 minute rest for a mile predictor. It doesn't always align with a race pace exactly but always gives the feeling of laps in a mile, which is the important thing to replicate.

Also, as with a lot of training, there is a big mental aspect. If you tell your athlete that she has just crushed a great predictor workout for 5k and should be due a great time, that builds confidence. Often a race can feel like a big occasion but if you are able to remind your athlete, during the race, that this is just what they ran in practice or just putting the pieces of that workout together, it can calm them and relax the chaos of race.

It seems like it is just a way to motivate athletes by putting some meaning into the workout. Often middle-distance runners will over-perform on predictor workouts for longer distances. It doesn't really matter - everyone still gets a good workout in.

Right, she is a 1500m / miler, a little aerobically underdeveloped but we have been turning the right stones the past month and I see alot of potential. Im not quite sure she could hold 3:32 / km pace for a 5000m race (yet) but its definitely a great session nonetheless without the reps being too hard or pushing into a highly anaerobic state. Just cruising through em!

Most predictor workouts are also good workouts, so they’re certainly not useless.

How well they line up with race results is going to be pretty individual, both in terms of physiology and how the athlete responds mentally to racing, but I suspect everyone sort of works out the more reliable indicators for themselves with time and ideally you’re using them in the context of other workouts/trends anyway.

Probably less necessary if you’re racing frequently, but it’s her first try at that distance having a ballpark figure and a rough idea if what the effort is going to feel like is probably a good thing- and even if it’s a bit off the mark, it’s still a data point that can help identify strengths/weaknesses and inform your approach in the future, which is what you’re after with those first couple of attempts IMO.

You answered your own question.

If I did 5 x 1k on 60s rest, I would expect to race a 5k faster than that. 6 x 1k w./ 90s-2:00 would put me right on pace.

Prediction workouts are usually good workouts. But some people take them out of context & overestimate fitness. 1 workout shows potential. A bunch of workouts strung together over a several week build tells you what you'll run.

It sounds like the athlete you're coaching could run a (sub) 17:30 5k pretty soon with some additional strength work and a sound pacing strategy.

If it's a good race-specific workout that will give your athlete confidence, then sure, go for it. If there is potential for a prediction workout to set an expectation that might end up in disappointment, I would maybe not do it like that, or at least frame it differently (this probably depends mostly on the athlete's personality).

I didn’t race at all last year for 6 months and gained a lot of fitness during that time. Started at about 16:40 shape and knew I was much much quicker than that by the end. I had a 5k race but didn’t know what pace to run, so I’m the 3 weeks leading up I did 12 x 400m off 60s, then 5 x 1k off 2 mins.

This put me right on 15:30 pace and I went out at that pace in the 5k and ran 15:33. So yeah it’s pretty accurate if you do it right