Good luck with it. I tried that method for about 5yrs. with mixed results. Works best with sprint races from 1000m to 1100m with fields of more than 7 horses [10 to 14 ideal].

The big problem in evaluating race times for races 1200m and over is race pace. Next run for any horse might be a slowly run race or truly run race. That's going to depend on the make up of the field. Take the Orr [G1]. The first 4 in r1 on that day all ran faster times than the winner of the G1. Pace can make times a bit meaningless except in short sprints where the foots on the accelerator from the jump.
Time becomes a bit of problem as an indicator of performance if a race isn't truly run [70% of races not truly run, would be a good guess for Eastern States]. WA seems to have the most truly run races [within 2sec of race record].

Anyhow, good luck with it.