The following email arrived at TurtleTrader. It describes a style of trading called scale trading. Our comments are in bold:

I was reading losers average losers on your site and it spurned the following query. A colleague of mine has started using a system that does just this, averages down. His strategy is to buy mini-index futures contracts, say 10 contracts for example, every x pts drop, lets say 100. He then sells those same 10 of contracts on a x pts rise. It sounded like suicide to me, until I actually ran it on a spreadsheet using several years data from 1997 to today. It gives massive drawdowns, but due to the dollar cost averaging effect, actually ends up making money over the long term [TurtleTrader® comment: This sounds like Enron revisited]. Lots of it – as long as you have enough capital to survive the inevitable drawdowns that come everytime the market trends down. It seems to be the antithesis of Trend Following, complete idiocy, against every rule I ever read in a trading book, and certain bankruptcy, but it makes money. If a market goes all the way down [TurtleTrader® comment: Then short it], and all the way back up [TurtleTrader® comment: Who says it will!], he makes 100 points for every contract he bought and ends up with no contracts. Until the market drops again, and he starts buying. If the market goes down and stays there, oscillating up and down for months, he is sitting on a potentially huge loss [TurtleTrader® comment: This is not potential, it is real!], but turning over contracts for every 100 point rise and fall that serve to reduce his overall loss and average cost of his contracts. The longer this oscillation goes on, the better off he becomes. At some point the market will do one of three things:

1) Go down again. If he has enough money (and he should if he made his initial calculations correctly, or assumed the index COULD fall to zero) he keeps playing, with an even bigger loss, waiting for the recuperation that will come when the down trend reverses. [TurtleTrader® comment: This is prediction. Impossible goal]

2) Stabilizes at a lower level and oscillates for days/weeks/months. He keeps playing, buying and selling every 100 points reducing his average cost. If this goes on long enough, the (high) initial average cost of the contracts he bought higher up will drop to a much lower level. [TurtleTrader® comment: This is prediction. Impossible goal]

3) The market goes back up again allowing him to sell some of/all of his contracts. Each one at a 100 point profit. Not an ideal solution, but nevertheless a profitable one. [TurtleTrader® comment: This is prediction. Impossible goal]

His rules follow. I’d be interested in your Trend Following opinion as it seems to do exactly the opposite of everything you talk about, and as I am a regular reader of your site and trying to either devise my own system or decide to buy your methodology.

1) His first rule is to make sure he has enough capital to survive the worst possible drawdown, which theoretically is zero. He combines position sizing and game theory to find the ideal number of contracts and point drop to use. As he trades an index, his belief is that, assuming there is no global meltdown, an index will never drop to zero [TurtleTrader® comment: This can never be known], as could a share, as poor performing companies are replaced by better ones, thus keeping the price of the index from reaching zero. Armageddon may happen, but he’d lose his money even if it were in the bank then.

2) Once he has decided what the most probable, worst case, minimum level for the market is, he divides his capital up into bets, with enough to cover one bet for every x pts drop as far as his worst case price, plus enough to cover the daily adjustments. For example, if the market is 10,000 and he trades every 100 points, he must have enough capital to make at least 100 bets and withstand the drawdowns from the daily adjustments on each contract he holds, from the point he bought it, to the current price (anywhere between zero and buying price) This is not a small amount of money, but assuming he has it, that’s the plan.

3) On the first 100 point drop, he buys his 10 contracts (the number of contracts is dependant on his capital, the current value of the market and the worst case price he calculates, which could be zero). If it falls another 100, he buys again, if it falls again, he buys again. When it rises 100 points he sells. This sale reduces the average cost of the contracts he holds. He may be sitting on an enormous drawdown due to a sustained downward trend from a high value, but as the market oscillates, each rise and fall of 100 points serves to reduce his average cost.

4) Obviously, at some point he will be sitting on a massive paper loss, but as long as he still has capital to pay his daily adjustments, he will keep playing [TurtleTrader® comment: But what if he runs out of money?]. Each oscillation serving to reduce the average price of the contracts he holds. If the market goes back up, he makes money. If the market stays where it is and whips, he will be drip feeding his account until either the market drops again (he keeps buying) or goes up, where he starts selling his 10 contracts every 100 points, until he has no contracts left. Its actually better for him that the market go sideways in an undefined way, just going up and down with no real direction. This allows him to churn contracts and lower the average cost. As far as I can see, this goes against everything you teach, although it does seem to make sense and, mathematically, assuming he makes his calculations correctly and his capital is enough to trade fixed lots all the way to zero and back up again, it seems to be a winning long-term strategy? [TurtleTrader® comment: It is a sure-fire way to the poor house. There is no money management or stop usage] I’m interested in what a Trend Follower would say (I think I can already imagine, but as you guys seem pretty clued up and on the ball, I’d be interested criticism about this system, combining position sizing and systematic dollar cost averaging. It seems like the road to ruin to me, but putting the numbers into a spreadsheet produces profits. [TurtleTrader® comment: Numbers in a spreadsheet? Can you spell Enron? Scale trading continues to buy as a market goes down. The hope is that the market will turn at some point and you will have a big position averaged at a lower price. Problems? You have to sit on losses for a long time. It may also take years for a market to recover (if at all). More? You may have to continually add money to your trading account for margin. And if the market continues to move lower, you could run out of money. This is a real bad strategy]

Why Scale Trading Fails in Practice

The reader’s instinct that the strategy “sounds like suicide” is correct. The spreadsheet that produces profits is the problem, not the solution. A backtest that shows scale trading working over a specific period of market history is a backtest that selected a period where the market eventually recovered. It is survivorship bias applied to time windows rather than to companies. The Nikkei from 1989 to 2003 is the scale trader’s nightmare: an index that lost 75% of its value and did not recover for decades. A scale trader who began buying the Nikkei in 1990, confident that an index cannot fall to zero, would have run out of capital long before the market turned.

Every TurtleTrader® comment in this exchange points to the same structural failure: the strategy is built entirely on predictions that cannot be made. “The market will recover.” “The index will not fall to zero.” “I have enough capital for the worst case.” Each of these is a forecast of an unknowable future event. Scale trading is not a systematic approach. It is hope with a spreadsheet attached.

Trend following’s structural opposite is not accidental. Where scale trading buys more as price falls, trend following cuts losses when price falls. Where scale trading has no stop loss and no defined exit, trend following defines the exit before the position is entered. Where scale trading requires the market to eventually recover, trend following requires only that price moves in a direction long enough to produce a profit and then follows it out. One approach is built on prediction. The other is built on rules. One approach runs out of money when the prediction fails. The other limits each loss to a predefined amount regardless of what the market does.

Frequently Asked Questions

What is scale trading and why is it dangerous?

Scale trading is a strategy of buying more of an asset as its price falls, with the intention of selling at a profit when the price recovers. It sounds logical in backtests because most markets eventually recover from most declines. It fails in practice because it requires the trader to hold massive unrealized losses through extended downtrends, continuously add capital for margin requirements, and correctly predict that the market will eventually recover before the trader runs out of money. All three requirements can fail simultaneously.

Why does a spreadsheet showing scale trading profits not mean the strategy works?

Because the backtest selects time windows where the market eventually recovered. Any period that includes a sustained decline followed by recovery will show scale trading working. The period that shows it failing is the one where the market declined severely and did not recover within the trader’s capital horizon. Japan’s equity market from 1989 to 2003 is the real-world example that any scale trading backtest should be required to survive.

What is the fundamental difference between scale trading and trend following?

Trend following cuts losses when price moves against the position and holds winners when price moves in the position’s direction. Scale trading adds to losses when price moves against the position and takes small profits when price briefly reverses. One approach limits the maximum loss by design. The other has no limit on losses except the trader’s capital. One approach is built on rules. The other is built on the prediction that price will eventually recover.

Trend Following Systems
Want to learn more and start trading trend following systems? Start here.