Time is money: Costing the impact of duration misperception in market prices

Abstract

We explore whether, and to what extent, traders in a real world financial market, where participants' judgements are reportedly well calibrated, are subject to duration misperception. To achieve this, we examine duration misperception in the horserace betting market. We develop a two-stage algorithm to predict horses' winning probabilities that account for a duration-related factor that is known to affect horses' winning prospects. The algorithm adapts survival analysis and combines it with the conditional logit model. Using a dataset of 4736 horseraces and the lifetime career statistics of the 53295 horses running in these races, we demonstrate that prices fail to discount fully information related to duration since a horse's last win. We show that this failure is extremely costly, since a betting strategy based on the predictions arising from the model shows substantial profits (932.5% and 16.27%, with and without reinvestment of winnings, respectively). We discuss the important implications of duration neglect in the wider economy. Highlights: Demonstrate existence of duration misperception in a real-world environment  Develop a 2-stage forecasting model combining survival analysis & conditional logit  Employ survival analysis to study pricing efficiency in a speculative market  Evaluate the economic costs of duration misperception in a financial market

Let's Call it Quits: Break‐Even Effects in the Decision to Stop Taking Risks

“…Consequently, we adopted the extended Cox proportional hazard (CPH) model with time‐varying covariates (Cox, ), a powerful multiplicative regression SA model that avoids potentially biased prior distribution assumptions and accounts for multiple risk factors simultaneously (i.e., covariates). It has been successfully adopted in many domains (Lane, Looney, & Wansley, ; Ma, Tang, McGroarty, Sung, & Johnson, ). Full details of the CPH model we employ are provided in the Appendix.…”

Improving prediction market forecasts by detecting and correcting possible over-reaction to price movements

“…However, it has been shown that bettors do not fully account for the interactions between multiple performance-related factors e.g., various aspects of a player's previous performances: see Sung & Johnson, 2008a, for survey). Consequently, it is possible to develop superior forecasts which combine and capture complex non-linear relationships between odds from parallel markets or odds and a range of variables associated with multiple performance-related factors (e.g., Ma et al, 2016;Lessmann et al, 2009Lessmann et al, , 2010Lessmann et al, , 2012Wunderlich & Memmert, 2016;Goddard, 2013;Dixon & Pope, 2004;Spann & Skiera, 2009;Goddard, 2005;Donniger, 2014).…”

To what extent can new web-based technology improve forecasts? Assessing the economic value of information derived from Virtual Globes and its rate of diffusion in a financial market

“…Equally, some models that combine a range of variables related to horse or jockey performance, capturing nonlinear relationships, can produce probability estimates that are not fully discounted in odds (e.g., Lessmann, 2009Lessmann, , 2010Lessmann, , 2012. Equally, some variables based on publicly available information derived from complex modelling procedures, are also not fully accounted for in odds (e.g., horses' recovery times/duration between wins: Ma et al, 2016;starting position: Johnson et al, 2010a; performance against common opponents: Knottenbelt, 2012). Importantly, with an increase in complexity, bettors may not be able to scrutinise fully all important relationships between variables and performance outcomes and feedback may become more ambiguous (Brehmer and Allard, 1991).…”