Currency forecasting: An investigation of extrapolative judgement

Mary E. Wilkie-Thomson*, Dilek Önkal-Atay, Andrew C. Pollock

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

9 Citations (Scopus)


This paper aims to explore the potential effects of trend type, noise and forecast horizon on experts' and novices' probabilistic forecasts. The subjects made forecasts over six time horizons from simulated monthly currency series based on a random walk, with zero, constant and stochastic drift, at two noise levels. The difference between the Mean Absolute Probability Score of each participant and an AR(1) model was used to evaluate performance. The results showed that the experts performed better than the novices, although worse than the model except in the case of zero drift series. No clear expertise effects occurred over horizons, albeit subjects' performance relative to the model improved as the horizon increased. Possible explanations are offered and some suggestions for future research are outlined.

Original languageEnglish
Pages (from-to)509-526
Number of pages18
JournalInternational Journal of Forecasting
Issue number4
Publication statusPublished - Dec 1997
Externally publishedYes


Dive into the research topics of 'Currency forecasting: An investigation of extrapolative judgement'. Together they form a unique fingerprint.

Cite this