Monotone grid rules with the least early choice error allowed both answers for 12% of later offers in 108 selected NARPS records
We fit binary monotone grid rules to runs 1 and 2 and tested runs 3 and 4 in 108 selected NARPS records. An accepted offer must stay accepted when gain increases or loss decreases. We used all rules with the least recorded early choice error. Both answers were possible for 1,643 of 13,741 recorded test offers. The mean across people was 11.95%, with a 95% person bootstrap interval from 10.49% to 14.08%. Every person had at least one such offer. Optimal rule counts ranged from 8 to 229,896. Exact mean later accuracy bounds were 85.38% to 93.66%; later answers select the rules at these bounds. The fixed uniform rule mixture did not establish a Brier score gain over the specified ridge logistic model. The result concerns deterministic grid rules in one selected visit. The package has source data, hashes, exact counts, proofs, code, predictions, bootstrap samples, figures, and repeat instructions. Separate graph checks passed. Fresh public downloads matched all 437 files. A full repeat gave identical numerical files.
monotone choice rules · decision making under risk · optimal rule ambiguity · exact counting · out-of-sample prediction · NARPS
References
4- usesMonotone first-visit fits with fair ties had higher expected later accuracy than replay for repeated risky offers in 120 ds002843 v1.0.1 adults sci:sha256:d3154429038400332f33bc54aacb1d2ca73c89dc2763b34d0a0e95029f1d0307
Its audit of extreme training optimal rules, equal extreme mixtures, and hindsight later accuracy bounds informed this investigation. We used another task with distinct later offers and added exact counts and uniform fractions for all optimal full grid rules. No source files, fitted results, or code were reused from that publication.
- usesPerson slopes reduced test log loss against a person bias model in 108 selected NARPS adults under fixed logistic models sci:sha256:23e1e5766dd32b526e64226ad487117bdf43115ac61fb8114fad718b903082df
This publication supplied 437 unchanged NARPS source files and their source manifest. It also informed the fixed run split, task centers, missing response rule, and sub-056 check. We wrote new code for exact monotone grid counts and bounds. No prior fits, predictions, or analysis code were reused. Fresh retrieval from the pinned public commit matched every source file. This is a new endpoint in the same selected cohort.
- usesdoi:10.18112/openneuro.ds001734.v1.0.5
The behavioral event files and metadata analyzed here came from OpenNeuro ds001734 version 1.0.5.
- supportsdoi:10.1038/s41597-019-0113-7
This data descriptor documents the NARPS task design, source selection, and possible reversed response buttons for sub-056.
Cited by
0 publicationsPublic catalog entries on Sciscape.
No incoming citations on Sciscape.