We know that, to become expert in some field, one has to do a lot of practice. 10,000 hours is the standard, popularised by Malcolm Gladwell but actually supported by academic research as well. And this should not just be any kind of practice, but deliberate practice, the kind where, essentially, you get decent feedback and try to improve.

Can you get the necessary practice in discretionary macro? Clearly not. You would have to spend 10,000 hours making macro calls on the markets. Even if we rather generously allow a whole 12-hour day for each decision — including the collation of information, the formation of an argument, the picking of an entry point and of course the actual execution — at my recent rate of about one decision a month it would take 70 years to accumulate the necessary practice. That doesn’t sound like a good proposition to me.

Many people think that discretionary macro trading (other than short-term, around data releases) is impossible. When conceived in the terms I have considered so far, they are right.

10,000 hours of deliberate practice are required in order to train system 1 in some discipline. But we are creatures of two systems, one automatic and one rational. System 2, the rational system, does not require training in a specific discipline in order to perform reliably (one needs years of training to use one’s reason effectively, but one does not need years of training for each and every question to which reason is applied). This does not tell us that macro trading is possible using system 2. What it tells us is that the above argument that it is impossible using system 1 does not imply that it is impossible altogether.

What would be the components of a macro trading process based on system 2? Here are some preliminary thoughts.

  1. The judgement about whether or not there is an entry point should be systematic; since my system 1 is inclined to want to trade when an opportunity presents itself (I have already trained myself not to be paralysed with fear), the system should be designed to prevent trading at inopportune moments.
  2. The judgement that there is a trade to be done should be absolutely explicit. It should be possible briefly to summarise the premises and conclusions in a way that makes a strong argument.
  3. The criteria for what counts as an argument should be well-defined in advance. There are two broad types of argument to be considered. First, there are arguments that some major factor ought to change and is already changing. Second, there are arguments that the market should not hit some particular level, at which the stop can be placed. These are separate and both should be required for a trade.
  4. You need system 1 — it is the motivator of action. But the action it should motivate should not be a particular trade, but the creation of an argument that counts as valid under the criteria.
  5. The knowledge that one requires to make a judgement should be easily available and conveniently presented. This ties in with the data sheets that I have been creating.
  6. It is foolish to gamble all of one’s success on a single entry point, however good it seems. One should scale into trades as appropriate entry points present themselves up to a maximum loss per trade.

I will keep working on this, but I want to emphasise that the main insight here is that one should not expect to be able to make good macro calls using “judgement” or “experience” — i.e. system 1 — before the age of 90 at the earliest. If system 1 cannot be used, one has to fall back on system 2. The only ways to use system 2 with any consistency are 1) no to use it to make particular decisions at all, but only to design a decision-making system and 2) to make arguments absolutely explicit and to hammer them from every angle to see if they break. It may be that only the former is appropriate for my kind of trading (a la Bridgewater), but I intend to attempt to combine the two for the time being.