In this case, if you changed your trial interval to say 200ms, then you could use "rt:57,full,clear400" with a min-max value of 0-1800! for the trial. You need some trial interval because that's when DirectRT prepares the next trial. If, for example, your standard trial interval is set to 200ms, then adding a total trial duration of 1800ms would produce 2000ms between trial onsets. You could also try a trial interval of 100ms and a min-max value of 0-1900! etc. Does that make sense? Note though that DirectRT will not accept a response during the trial interval, so in the latter case, you would be looking at 1900ms of response time potential.