Latency-probability curves for sequential decision models: A comment on Weatherburn