I post a lot of critical thoughts, and spend a lot of time encouraging people to think deeply about the technical tools they use. Recently, I’ve received a lot of questions about what I think actually works. Though I’ve answered this in some depth before (and have written a fairly large book on the subject), maybe a concise post here would be useful.
Conceptually, think of price action as being shaped by two conflicting forces: mean reversion (the tendency for large moves to be reversed) and momentum (the tendency for large moves to lead to further moves in the same direction). Most times in most markets these forces are in balance. When they are in balance, price movement is nearly random. Over any large period of price movement, these force will more or less balance, which is why academic studies still find good support for random walks in market prices. It’s also why traders have to wait for their spots, and why you can’t simply trade any market at any time.
Are there patterns that can show us if one force or the other is likely to dominate price action over a certain timeframe? That’s one way to phrase the essential question that has motivated all of my research and thinking about markets, and it’s also one way to think about the problem of technical trading. Can we find patterns that show us when we should be “going with” large moves or when we should be “fading” those moves? In the absence of such patterns (or if they do not exist and market action is random), then we’re just trading in randomness and it is impossible to make money.
The good news is yes, it is possible to identify these patterns, and they are not complicated. Yes, there is room for great subtlety and refinement in application, but, at their core, these are simple price patterns. Here are a few examples of price patterns that can tilt the scales in favor of one force or the other (not an exhaustive list):
Mean reversion
- Fading large single bars
- Fading N-day runs
- Fading breakouts of N-day highs or lows
- Fading large excursions from average prices (think Keltner channels or Bollinger Bands)
Momentum
- Pullbacks
- Nested pulbacks
- Breakouts, in some cases
- Volatility compression
In my experience, this is all there is: the understanding that these two forces shape prices, they are usually in balance, and there is usually nothing to be done in any particular market because there is no edge. The game, then, is redefined as waiting for a pattern that shows there is a potential imbalance, taking a position (the correct position, with the correct risk) to capitalize on the imbalance, and managing the trade as it develops. That is all there is, and it’s enough.
Love this one!
Pingback: The Whole Street’s Daily Wrap for 10/23/2014 | The Whole Street
This is like the Subtle Difference between Simple and Easy …. One Day !
Digging deeper into technical trading it has already become clearer, after a few days of book reading en googling I’m starting to see some patterns in between the randomness of a stock market, now all I need more study and a bucket load of practice, I’m a true believer that practice makes skill, not just shear brainpower, but it doest take time.
Is it common to dream about stock in your opinion? 🙂
Great post, thanks!
Pingback: ES SP500 Spoo-nalysis (Page 930) - Index Futures Trading | Big Mike Trading
Pingback: Random Line Theory (Page 25) - Traders Hideout | Big Mike Trading
Adam, I’ve never heard it put so succinctly. May that be the forward to your next book! It also helps me understand why the academic view of random walk theory can hold, yet we still trade. You help to clarify this grey area and a topic which normally brings so many ‘Us Vs Them’ arguments.
Thank you. I will have a post on that very topic soon too.
With regards to momentum trading…what do you mean by pullbacks? Do you mean buy the pullback (in the case of a long trade). So it is basically a counter trend entry and a trend based exit? Also, what is a nested pullback? thanks!
Look at the pullback topic on this blog and you’ll see many examples. Yes, you are correct. I also think there are some nested pullback examples, which is a smaller pullback in context of a bigger one working toward its target.
“The game, then, is redefined as waiting for a pattern that shows there is a potential imbalance”
Which will last for how long? One millisecond, one second or one minute?
IMO, “imbalances” is the wrong way of looking at this. What you call imbalance is balance for someone else. It depend son the Point of views since buyers always match sellers.
IMHO you are right insofar as there might be an imbalance on one timeframe and balance on another, but this does not dismiss the concept. But what is your way looking at this?
Well, I would take issue with “buyers always match sellers”. What does that mean? That there are an equal number of buyers and sellers? That’s false–the old “there’s a buyer for every seller argument” is simply wrong. Certainly there are cases where people want to buy a market with more conviction than people want to sell it (and vice versa). This is an imbalance, and this is why markets move from one price level to another.
I still hold that the quote you began your comment with is literally the single most important thing a technical trader can learn.
I think both of you can be right. Let me clarify:
If an actual trade has been filled, buying and selling would have been matched. But there can be still an imbalance of potential or unfilled (!) buying and selling which can manifest ifself in a kind of buying or selling pressure which finally drives prices out of balance.
IMO, perhaps the most important point Adam makes in his writings on “what works” is that there is no singular “market”; asset classes often behave wildly differently to one another. Most TA “educators” would like you to believe that their magic bullet works equally well between ES, CL, 6B, etc. but if one runs their own tests, they’d find that prices move *very* differently in each of these markets.
For me, this is a very inconvenient truth in that it makes statistically significant results more difficult to achieve (particularly on higher timeframes), and also creates logical challenges when looking at data (e.g. it isn’t hard to imagine equity and commodity markets behaving differently because the underlying assets are so different in nature, but what if a system test returns statistically significant results in crude, but totally random results in natural gas?). But, hey, who ever said trading was easy?