Ed Seykota was using moving averages crossover
He was.
He was using 80-period and 140-period exponential moving averages of daily stock closing-prices as a method of trend-identification
for the purpose of directional bias only.
He was.
Note that he was
NOT using moving averages
crossovers as a basis for trade entries.
Not sure if he is still using the same system.
He isn't.
But clearly he did well with it,
as a timeframe-specific, trend-based, bias-determination method, for a while, and unlike some of the people posting here, his results (obviously) were statistically significant.
I won’t try to defend my position
You don't
have a "position" to defend.
What you have are the results of 13 consecutive trades, of
no statistical significance whatsoever.
When you have the results from 13,
000 (or perhaps even just 1,3
00) trades, showing the same thing,
then you'll have "a position to defend" (and it will be an interesting one to defend, because it will conflict with the findings of just about every independent, objective, systematic, statistically significant research-trial which has ever been published on this subject!
).