The Pseudo Trader
by Mark Harris
How can you model stocks without stocks? By using random numbers to drive the simulation. In
previous articles (STOCKS & COMMODITIES, December 1990 and January 1991), first I demonstrated the
modeling of stock charts using this technique and then the modeling of a stock specialist. The problem I
encountered was that the model lacked feedback — while volume influenced price (via the actions of the
pseudo specialist), the reverse did not hold.
What happens when the loop is closed by adding an object that responds to price? I call this new
Smalltalk object a "trader," for reasons that are fairly obvious. Similar to a real-life technical trader, the
goal of this object is to profit by applying some system or formula to the reported price series.
Visualize this pseudo trader in two ways. First, the "trader" might literally be an individual with very
deep pockets. Second, the view I prefer, we can think of the "trader" as the concerted action of a number
of individual traders who just happen to be following the same system. With this second interpretation
we can get a feeling for what happens when a particular system gains more adherents.
I have not created a day trader in this simulation; rather, I have created a trader with access to closing
data (high, low, close) for a given day and, on the basis of that data, the trader can act in time to catch
What should be done when the order is larger than the specialist is willing to handle? The pseudo
specialist will not handle orders that place him over his capital limit. The mechanism in the model that
deals with this simply places over the limit orders on the book, where they are held until the random flow
of orders allows them to be filled.
The pseudo trader, on the other hand, isn't willing to place his (large) order as a market order that may be
executed in bits and pieces at successively higher (or lower, if selling) prices.