So i'm trying to think up a small field experiment using the Duel arena in OSRS. I'm trying to investigate a behavioural effect - impact of scaling on decision making.
I've collected data from every whip staker, their in game wealth, and the proportion of their budget that they've staked. Filtered for specifications etc.
I've ordered the data by budget and then split into decile segments. (Top and bottom 2 deciles are excluded as per spec)
I've approximated a probability density function for each budget band(KDE). (Probability density of budget proportion staked)
I've done a K-sample of Kolmogorov-Smirnov test to test whether the distributions are different (ie whether or not the effect i'm trying to test EXISTS)
I'm not sure how I could go about measuring how the effect works though IE distance. At the moment I'm thinking of approximating wasserstein distance between each segment? Could I even infer anything useful for that? Is there a simpler approach that i'm missing?
I've collected data from every whip staker, their in game wealth, and the proportion of their budget that they've staked. Filtered for specifications etc.
I've ordered the data by budget and then split into decile segments. (Top and bottom 2 deciles are excluded as per spec)
I've approximated a probability density function for each budget band(KDE). (Probability density of budget proportion staked)
I've done a K-sample of Kolmogorov-Smirnov test to test whether the distributions are different (ie whether or not the effect i'm trying to test EXISTS)
I'm not sure how I could go about measuring how the effect works though IE distance. At the moment I'm thinking of approximating wasserstein distance between each segment? Could I even infer anything useful for that? Is there a simpler approach that i'm missing?
