I hate to dump on that article but he does what many do when they try to undermine "market timing" or disaster avoidance techniques. Nobody, but nobody, anywhere, at any time goes in and out of the market based simply and solely on the basis of a closing price above or below the 200 DMA. There is always some kind of capacitor involved to avoid the kinds of return-sapping whipsaws he talks about.

1) Waiting for a close x% above or below the 200 DMA
2) Waiting for X number of days in a row closing above/below
Something like that.

The only strictly binary signal I know of uses the 10 month MA and that was shown to be, over longer periods of time, as good or slightly better than B&H. I think it was Mebane Faber who did the work on it.