Closed DrPaulBrewer closed 4 years ago
The sampler was used for the first, incorrect figure and is associated with this issue.
The period filter seems to work ok and produced the upper triangular expected result.
The extract()
call handles sampling and filtering.
Extract is called twice, once for each element of chart.names
(the x-axis variable and the y-axis variable)
Instead, what is needed is to call extract once, and transpluck to get related x and y variables instead of unrelated.
Thoughts:
transpluck(source, {pluck: [var1,var2,var3]})
can extract multiple variables as an object of arrays.Confirmed fixed in 6.0.1. Got expected triangular 2D histogram for trade values/cost.
Ran 100,000 periods of a market populated by "Truthful" agents.
The 2D Histrogram for trades, buyerValue vs. sellerCost is showing a positive z value for cases where cost exceeds value.
The histogram can be seen here
Trades where sellerCost exceeds buyerValue should not occur.
Because of this, the "trades.csv" file was isolated and loaded into an sqlite3 database to search for these trades.
However, the log/database shows no trades where cost exceeds value -- as expected.
zip file: 20200720T064506.zip
the contained file: trades.csv
was loaded into sqlite3