Sep

13

 Something I've been considering lately is how to combine different types of consideration that occur in different time frames. This seems to be an issue in both chess and speculation, for example stocks have a long term bullish bias which short term traders may sometimes forget. Chess meanwhile has thinking based largely on the calculation of variations, and this needs to be married with long term almost glacial changes involving the pawn structure.

The main issue with chess at least seems to be that the longer time frame tends to be sidelined or ignored, particularly in games in which it's 'all hands on deck'. So players often calculate like mad and lose sight of the 'big picture', often to their cost.

One thing that I've found can help is to set aside particular times for the different thought processes. Alexander Kotov, in 'Think Like A Grandmaster', reported that Mikhail Botvinnik tended to do this in his own games, considering 'strategy' (pawn structure) when it was his opponent's turn to move and then calculating variations when it was his own. I think this is a useful habit to cultivate, though it can be derailed if an opponent moves particularly quickly after a structural change has occurred.

It would be interesting to hear if any specs have comparable examples and how they organize their thinking to cope. Is there a decision tree, a simple heuristic ('trade smaller on the short side') or does everything just fuse together when enough experience has been gained?

Let's say that most fed/economy/war issues are going to develop over weeks and months and that we take a particular view (eg that QE will continue as and when needed until systemic failure). How does one then marry this with shorter term considerations, such as the Friday-Monday effect?

Anonymous writes:

Nigel, I think this is actually much easier in the markets than in chess.

Firstly, in chess, you are forced to make a next move at the lowest time unit. The clock is ticking literally. In the markets, you can always walk away from the table and choose to play at a completely different game where the risk/reward is more appealing.

Secondly, in the markets, it is very easy to implement a low pass or high pass filter. The simplest example is to use only weekly or monthly prices with no high/lows. Other examples of filters to allow one to see longer time frames including moving averages. Personally, I use multiple strategies at multiple time frames — and some of the strategies are canceling each other at any given moment. This requires extreme mental discipline to implement and maintain. But it results in a very different overall performance than focusing on a single timeframe.

Thirdly, in the markets, the game never ends. At the checkmate or resignation or draw, the game has ended. This end of game phenomenon directly affects risk taking choices and time horizon. I would hypothesize that the issue you raise is more similar to risk taking on games of chance like horse races (and the concomitant Kelly strategy) than on continuous outcomes.

Lastly, and most importantly, unless you are managing a massive amount of capital, one's "move" in the markets does not change the outcome. In Chess, your risk taking and behavior directly affects the future.

Jeff Rollert writes: 

Doesn't the "game" end when you close out a position? I could see the game continuing if you repeated a strategy, but the probabilities would be different, hence a different game.


Comments

Name

Email

Website

Speak your mind

1 Comment so far

  1. Edward Lam on September 10, 2015 5:08 am

    I definitely think that in most contexts it is better thinking about one thing at a time:
    http://www.joelonsoftware.com/articles/fog0000000022.html
    However, as suggested by “life imitates chess” (Kasparov) perhaps, the last 5% of value lies in “fantasy” in day dreaming combinations - which I see as a form of “multi-processing” (though not necessarily multitasking…

Archives

Resources & Links

Search