Changes in the runnup are twice as big of a problem if you are trying to analyze fractions and pace. You can't project pace with a reasonable degree of accuracy like you can with final time because horses rate.

I use a multi-faceted approach.

How many front runners were there and how rateable were they?

Did the frontrunners move away from the rest of the pack early or was everyone bunched up?

Were the jockies urging the frontrunners or were they wrapped up on the lead?

If there were other races at the distance were the fractions extreme in either direction relative to the final time?

Did several horses that were on or near the lead (especially the 2nd and 3rd speeds) run lower final speed figures than expected?

I try to form an opinion based on all the evidence.