Following Ivan, it seems that we need to transform raw signal -1/+1 into 0/1 in order to use tradeSignal.m accurately. suppose pop=[1 0 0 0 1 1], tradeSignal(pop,[1 -1])=0 while tradeSignal(pop, [1 0])=1, the latter being the correct signal given the rule.
I do not think there is a forward bias, but it indeed is a kind of curve-fitting based on historical data. I did simulated trading based on this strategy on a daily and intra-day basis, and it has been a losing one.
Expecting to see the use of Johansen procedure in the determination of cointegrating relationship in a multiple-assets stat arb trading environment. The pairs trading part of the webinar still focused on two series and EG -- no big change from previous webinar.