Quality Money Management : Process Engineering and Best Practices for Systematic Trading and Investment

(Michael S) #1

259


Let us assume that a trading system performed 100 trades, of which 47 were profitable,
33 were unprofitable, and 20 were scratches. If d is the number of defective, or unprofit-
able, trades, then p  d / n is called the fraction defective. If the number of unprofit able
trades d  n. p is more than expected or traditionally acceptable, the process is out of
control. The acceptable limits for d can be shown on a control chart. Like control charts
for variables, a d chart can be constructed from given standards obtained from the backtest.
The control limits for d are

UCL
LCL













dd
dd

3


3


Upper and lower control limits (UCL and LCL) represent bands three times the stan-
dard deviation above and below the mean of a variable. Upper and lower specification
limits (USL and LSL) are bands above and below the mean which define acceptable
performance of a process.
Sometimes the fraction defective, p , is controlled. The control limits for p can be
easily derived:

UCL


LCL














p

pp
n

p

pp
n

3


1


3


1


()


()


respectively, where p  is the mean of the fraction defective for the whole population.
d and p charts are both useful, but the p chart is more robust when different sample sizes
are used.

28.3.3. Other Charts


There are many other statistics-based tools that organizations can employ including:

● SPC control for other known distributions.
● Pareto charts to separate the important few factors from the trivial many.
● Check sheets to display collected data in a manner that is easily interpreted.
● Histograms to chart the frequency of occurrence. Looking at the pattern of a histo-
gram will tell you:


  • Whether the process is normally distributed.

  • Where the process is centered.
    ● Scatter diagrams to determine correlation between variables.
    ● Run charts to display process results over time.
    ● Control charts (SPC) to determine whether process variation is the result of the
    process ’ s inherent capability or caused by some controllable external factor.
    ● Cause and effect diagrams.
    ● Stratification.
    ● Scatter diagrams.


28.3. SPC CONTROL CHARTS
Free download pdf