nucleation

LIT-1062: Predicting Short-Term Remembering as Boundedly Optimal Strategy Choice

2 min read

Properties4
StatusActive — read in full 2026-10-15 from the text of the publisher's typeset article as deposited in the University of Michigan's Deep Blue repository (hdl.handle.net/2027.42/133633), served as extracted full text by CORE's reader page for that record (core.ac.uk/reader/86638046); Wiley, the Birmingham and Deep Blue PDFs, and the Internet Archive's captures could not be fetched from this environment, so figures were not viewed. Worth reading as the individual-level test of bounded optimality in short-term remembering: after a forced phase measures each person's copying rate at every list length from 3 to 9, the length with the highest measured rate predicts that person's later choice better than one fewer or one more (55% vs 8% and 17% of choices, N = 40; 46% vs 12% and 14%, N = 20), an exponential reward for long lists raises the chosen length, and raising the cost of an error to the whole trial's points does not. The prediction is each person's empirical argmax, not a model: there is no capacity or look-up-time parameter, and the task has no look-up at all, since names are shown once and cannot be consulted again, so the alternative to holding more is more, shorter trials. For QUESTION-095 it shows people can find the best point of a memory-versus-overhead trade by rate, and adds one term Gray et al. 2006 (LIT-1073) lacked, a time cost for a wrongly remembered item; it has no term for query predictability, external retrieval or search, staleness, reuse across trials or erosion of the alternative, and it corrects the description of it, from Lieder and Griffiths (LIT-1071), as a memorize-or-look-up trade priced by look-up time.
Tagscognition
Cited as “paper” byNOTE-882 — Howes, Duggan, Kalidindi, Tseng & Lewis — Short-term remembering as boundedly optimal strategy choice
Cited as “source” byTHEORY-330 — After a forced phase in which each person experienced every memory load from three to nine items, the load that had given that person the best reward rate predicted their later choices better than one item fewer or one more, in a copying task where items are seen once and cannot be looked up again; a steeper reward for long lists raised chosen load, a higher price for errors did not, and about one person in five kept choosing a non-optimal load

Graph View

Howes, A., Duggan, G. B., Kalidindi, K., Tseng, Y.-C. and Lewis, R. L. (2016). Predicting short-term remembering as boundedly optimal strategy choice. Cognitive Science 40(5), 1192–1223. DOI 10.1111/cogs.12271

Key takeaways

Read in full (NOTE-882). Participants copied names one at a time from a mock email to a mock calendar and, after a forced phase that measured their own copying rate at every list length from 3 to 9, chose how many names to take per trial; the list length that had given each person the highest rate in their own data predicted their choice better than one item fewer or one more (55% vs 8% and 17% of choices in Experiment 1, N = 40; 46% vs 12% and 14% in Experiment 2, N = 20). An exponential points schedule raised chosen list length (6.99 vs 5.07), but raising the cost of an error so that one wrong item voids the whole trial did not (5.00 vs 5.05). The “boundedly optimal” prediction is the argmax of each participant’s measured rate, not a model with capacity, error or look-up parameters, and the task has no look-up: names are shown once, in sequence, and cannot be consulted again. 13 of 60 participants persistently chose a non-optimal length, 10 of them one item more.

Why this is here

Lieder and Griffiths (LIT-1071, NOTE-873) name it as the one place their review sets internal storage against retrieval from the environment. Registered at the owner’s request for QUESTION-095, where it is a candidate cost model for the choice between remembering and looking up, beside Gray and Fu 2004 (LIT-1058) and Gray et al. 2006 (LIT-1073).

The reading (NOTE-882) corrects the description above, which followed Lieder and Griffiths. The task has no look-up: names are seen once and cannot be revisited, and the choice is how long a list to take per trial. There is no parametric model; the prediction for each person is the list length with the best reward rate in their own measured data. Capacity and look-up time are not parameters, and raising the cost of misremembering did not change choices. It is a trade of memory load against reward rate, not of memory against the world.

Backlinks