We made no model changes. We did do an update of the networks used to
calculate the forecast. We do this about three times a year. A library update
means that we remove some networks that have not been performing well and include
networks that have. The updated library trades about 4 times more often than
the previous library. Since we had only 1 trade in the past three months, I see
no problem here. The computed return with the updated library is about double
that of the previous library.