Does Any Technical Analysis Work? (Part 2)

The initial impressions of trading system development discussed last time begs the question about whether technical analysis (TA) is all hoax.

I did an internet search for “do simple technical indicators work” and found some interesting posts.* Here is the first:

     > There is no blanket strategy of “TA.” No TA gospel definitively says
     > buy/sell when the price crosses the 200-MA. You’re simply looking for some
     > pricing pattern that appears in the chart to influence your decision to buy
     > or sell, but there are literally an infinite number of pricing patterns.
     > Maybe it’s not a 200-MA but a ratio of 200 * (1 minus the index
     > increase over that period). People come up with all sorts of weird
     > algorithms; some people even find one that will work for a bit. Ultimately,
     > it could be anything… [it qualifies as TA] as long as it involves the
     > quoted price relative to some historical metric… [emphasis mine] [1]

This argues certain combinations of TA indicators can be developed into profitable trading strategies. Yay!

This also provides some hope for data mining, which is an approach I will discuss in future posts.

Here’s another take that is similar in nature:

     > You seem to think “TA…” is some fixed, specific thing… there are a
     > hundred different ideas on so-called “TA” and every person who trades
     > argues bitterly about different aspects of charting… everyone who
     > uses one form of charting assumes that every other charting technique
     > is completely stupid. [emphasis mine] [2]

Mention of “hundred” different TA ideas is deceptively small since the actual amount is infinite (see [1] above).

Unfortunately, people posting online engage in name calling, flaming, and trolling. I often see an oozing excess of pride, arrogance, and ego. In the world of statistics and valid trading system development where strategies either pass or fail, I see no reason to get emotional or to passionately defend things. Follow the data and start testing where the data ends.

Out of those commenting online about TA and trading, I wonder what percentage actually has supporting data? Nobody is credible without large sample sizes under their belt and that includes number of strategies tested.

My algorithmic trading journey has just begun with 300 strategies tested. Even were I to have tested 10x that, though:

     > …just because YOU haven’t found one that you think works well enough
     > to rely on… [doesn’t] mean… [others have]. And when people do, they don’t
     > divulge it. If someone had a strategy that worked, it would not be on sale. [3]

Similarly:

     > When people figure something out and it’s vital to their own profitability,
     > they don’t run around sharing it. Why don’t people understand this? It
     > would save a lot of heartaches… [4]

This is a very interesting concept to consider. In other words, don’t expect anything you see, hear, or read to actually work.

Gotta love the smear campaign, too. You can’t say online financial forums lack entertainment!

I’ll continue next time.

* — Online content may or may not be credible. You have to decide for yourself, and this applies to my blog as well.