Most of us are just trying to form an educated guess as to the supply / demand balance.

If a stock is going “down” (however you define it … lower lows, moving average declining, etc, in your chosen timeframe) … then most technicians will not buy it no matter how “cheap” it may appear on a fundamental basis. We will short the thing, it may well get cheaper.

“up” and “down” are empirical concepts (again, depending on metric and timeframe)

“cheap” is ill-defined and, like beauty, is in the eye of the beholder.

Doesn’t matter how cheap it is, you buy it to go up.

Not disparaging the folks who are good at fundamental analysis … in fact, I believe the best approach is a combination:

“cheap” and “going up” or

“expensive” and “going down” for us folks who like to short

I’ve seen more abuse of statistics than correct use in my days following Wall Street; if I am a critic on backtesting, it is because I have seen it misused. It has a genuine place in our toolbox if judiciously used.

Yep, over fitting is a big issue with back testing, but theirs several steps you can take to lessen its affects. Test with out of sample data. Check “brittleness” by varing parameters and checking performance. Also with any system you should have an idea why and how it works, and why it should keep working into the future.

That said its hard to respect the opinion of a technician that won’t back test, precisely because its so easy to do so. Technical analysis is great as long as its well tested.

As for fundamental analysis, as much as possible that should be subject to backtesting too. If you can get reliable data you can certainly run tests related to P/E, PEG, cash flows etc..

The problem come into play with quantifying things like competitive moats as you mentioned. Even worse, any future looking fundamentals are impossible to quantify. For example “How do you factor the rise of the internet on old media companies?”

Yes a lot of FAs out there aren’t as rigorous as they should be. That said, FA isn’t the mirror image of TA.

I have my criticisms of backtesting; I’ve had to do a lot of financial modeling over the years, and I am wary of specification searches that tease out the perfect past, with all of the aspects that made it unique, and don’t respond to the uniqueness of the future.

That’s why I try to stick to relatively well proven anomalies that are seemingly out of favor in my investing. I may not have done the backtest, but if multiple others have come to a reasonably similar result, and there is a logic behind it from a behavioral investing standpoint, I use it.

As an aside, my oldest son started paper trading with my eight rules, and was beating the averages regularly. I was happy to find that he was winning, but not beating me. That gave me some small hope that what I call my “business judgment” might carry some ROI.

There’s merit in testing all approaches. Many different approaches work, including many of the ones that I choose not to use, and many of the ones that I’ve never investigated. We should all use approaches that have been tested (preferably by US), have been shown to work, and that we are comfortable with.

They don’t backtest and can’t describe how their current “optimal” strategy would have performed in the past.

Some of the items they look for have no edge in testing, if they even could be tested (tried to quantify a “moat” for a computer program lately?).

All funnymental guys know is buying low and watching it go lower while they hold on and hope.

They might buy high if the valuation is cheap, maybe, that’s the only case.

Bottom line, if one of those fundamental guys gives you a piece of advice, aske them how it has backtested. If they give you a blank look, ignore them.

Most classical tech analysis has shown to have little to no edge in actual testing. These are things like bull flags, rising wedges, head and shoulders, etc.

And yeah most tech analysis is built around buying strength and selling weakness. Momentum or Buy high, sell low.

The only case where a tech might buy low would be at a support line, but usually they are looking for a break out over resistance.

Bottom line is if a tech gives you a piece of advice, ask them how it has backtested. (Avg drawdowns, sharpe ratio, Std. dev., RoR, etc.) If they give you a blank look, ignore them.