The choose-short effect and trace models of timing.

McMillan et al. (1999) built a math model of the choose-short bias. This is when pigeons pick the short-time key even when the real time was longer.

They asked: what happens if we light up the dark gaps between trials? They tested the idea with computer runs, not live birds.

Bright gaps cut the choose-short error. The bias did not vanish; it just shrank.

The model said the reason is less memory mess, not a wiped memory.

Staddon et al. (2002) took the same trace idea and added a one-back rule. Their tuned-trace model covers more timing facts, like scallops, that the 1999 paper left out.

Arantes et al. (2008) ran real pigeons and showed context alone can flip time choices. Their data back the claim that signaled gaps matter.

Parmenter (1999) warned that choice errors may come from cue mix-ups, not memory loss. E et al. answer: even if cues are clear, trace clutter still bends choices.

If you use fixed-interval or delay programs, mark the blackout or wait time with a clear signal. A simple lamp, click, or color can trim timing mistakes without new contingencies. Try it next session: signal the inter-trial pause and watch if responses line up better with the true interval.