I use standard (average) times. It's important though to work out a daily track variant if you can because if the track is running fast then this will be reflected in the times being run. Conversely, if the track is running slow: meaning that a time run on a slow running track surface can be better than one run on a fast running track even though the latter time is the quicker.

If too many horses @ a meeting are running times faster than the normal be careful, likewise if there are bunched finishes in such circumstances - time normally will & should separate the wheat from the chaff & when it doesn't it should sound an alarm bell - good horses win in slow & fast times - poor horses normally only win in slow or average times - well worth remembering - and when this isn't the case there is normally some other reason such as abnormally fast track conditions or wind factor.

If you are using times then it is also necessary to fit those times to a profile of how the race will be run - racemapping - it's of little use selecting the fastest horse in the race if the race is likely to be run in a slow pace & therefore the best attribute the horse has is somewhat diminished.

My own ratings are based more on this latter factor - pace - then just times alone as IMHO times should be used not so much as pure ratings but more of a tool in detecting future winners or over-rated performances.

Daniel O'Sullivan who used to contribute to this forum has written some very good stuff on this issue.