Previously I posted about order types and mentioned that there was a lot more to say about the topic of stop orders.  It turns out there’s a deep relationship between stop orders, trend following and false breakouts that goes a long ways towards explaining market behavior.

The first key point to understand is that trend following is always conceptually a stop order and stop orders are conceptually trend following.  They’re two sides of the same coin.  Recall the definition of a stop order:

stop order – a stop order is a triggered order, with the price of the order being a trigger price.  Once the stop price is hit, the order converts to a market order.

and the definition of trend following:

trend following – a family of trading strategies that enter long positions when price reaches historic highs or short positions at historic lows.  These positions are held as long as new highs/lows continue to be made, and are exited when the pattern breaks.

Clearly there’s a relationship –  after all a trading strategy is nothing but a means of triggering orders, and if a strategy triggers a buy order when a sufficiently high price is achieved that’s logically a stop order.  This is true even if the trend following strategy didn’t enter it’s order using the exchange stop order type and instead entered it as a market or marketable limit order after the trigger occurred.  In other words, stops can be entered directly in the market or “held” in a remote trading system, but they behave essentially the same either way.  Trend following entries are the same thing as stop orders.

Now, I want to convince you that stop orders, even those used to bound losses, are conceptually trend following.  Imagine you buy 100 lots of XYZ stock at 26.54 and place a protective stop to sell those 100 lots if the price falls to 25.00.  This may well have some attractive behaviors from a risk mitigation standpoint, and I don’t want to minimize those.  But it’s also trend following – you’re in essence betting that if price falls that far, it’s going to keep falling.  So you sell.  That’s the same bet a trend follower entering a short position via a stop at 25.00 is making.  Again, symmetry between stop orders and trend following.

Last post, I discussed that false breakouts are the market’s immune system against trend followers.  But I didn’t explain what a false breakout is or the exact market mechanics of how they work.  Now I want to remedy that.  False breakouts are simply the price manifestation of a market behavior called a “stop hunt”.

stop hunt – a multi-step trade designed to take advantage of the presence of stops in the market.  For sake of discussion assume the stops are SELL STOP orders below the current price (perhaps used as stop loss protection).  Step 1 of the trade is to enter a BUY LIMIT order below the location the stops.  Step 2 is to enter a SELL MARKET order down to the stop price, triggering the stops.  The result is that the stop triggers and converts into a market order which sells into your limit order, and ideally your limit and market order are exactly the same size and cancel each other out perfectly. This leaves the stop hunter with a flat position and a guaranteed profit.  This trade can be reversed to hunt BUY STOP orders above the current price.

The ability to actually conduct a successful stop hunt is based on three things:

1. Accurate estimation of the prices where stop orders are present and the cumulative order size.  Since stop orders are not visible as part of the order book this is a non-trivial problem.
2. An accurate count of how many contracts are needed to move price to the stop order(s).
3. There not being too many people ahead of your limit order FIFO-wise that get filled by the market orders the stop(s) generate before you do.

Items 2 and 3 can be deduced by simply looking at the order book and counting up the size of the orders.  But item 1 is tricky – in order to hunt stops effectively, you need to know what price they will be entered at.  This is where we circle back around to trend following because trend following systems have a tendency to enter stop orders at well defined prices – buy stops above historical highs and sell stops below historical lows.  While this information is not enough to guarantee a stop hunter that stops will be present at a certain price and quantity, it is a pretty good start.  In fact, it’s good enough that stop hunters will drive price to likely trend following entry points just to see if stops can be found.  This price movement, experienced from the perspective of a trend follower, is a false breakout.  Here is the price chart from the last post with some additional annotations:

These annotations are somewhat fanciful in that I don’t know for a fact that the reason behind each price move.  The color commentary is entirely my perception.  But none the less they illustrate the principal of the stop hunt – buy stops accumulate above historic highs and sell stops below historic lows.  There’s then an incentive to move price to those points to trigger the stops.  Once that happens, there’s no reason for price movement to continue and it frequently moves back the other direction.  You can see how this behavior frustrates attempts at trend following – there is eventually a real trend here, but before the trend follower can profit from it, they take losses due to each of the false breakouts.  The net result for the trend follower is likely about break-even minus transaction costs.

Incidentally, I didn’t cherry pick this chart for the purposes of this post.  It just happened to be what the 1-minute ES chart looked like a few mornings ago.  I was simply lucky that it so clearly illustrated the concept of the false breakout.