|
| Test
Parameter Summary |
| Stocks |
S&P
100 (100 stocks) and S&P 500 (500
stocks) |
| Test
Period |
September
1, 1999 through August 31, 2003 |
| Optimization |
TURNED
OFF |
| Commissions,
etc. |
$.01
per share
commission
(achievable from firms like
Interactive Discount Brokers)
+
$.01
per
share
slippage. |
Result
Summary
Shown below is a graph of Annual Rate of
Return measured for all trades in the Test
Period for the S&P 100, as Tradability
Index is increased from a low value of 40
to a high of 90.
Plot of Tradability vs. APR
for the S&P 100. As Tradability Index
is increased, so does APR. click
to enlarge
Each
point in the graph shows profitability of
all trades where TI was above a given threshold,
along with the number of trades generated
(next to each data point.) We note that
APR reached a high point of about 365% when
TI was set at a threshold of 90 and higher.
We also note that the quantity of trades
decreased, from a high of 126 trades where
TI was set at 40 to a low of 5 trades when
TI was increased to 90.
Results
for the S&P 500 are shown below:
Plot
of Annual Rate of Return for certain thresholds
of Tradability
Index for the stocks in the S&P 500.
click
to enlarge
Analysis
of the Results
The relationship between TI and APR was
nearly linear in this test, which tends
to prove that Tradability Index is a viable
analysis metric. We also observe that stocks
in the S&P 100 were inherently more
tradable than those in the S&P 500 list.
For this reason, the same graph for the
S&P 500 is lower (about 105% APR for
TI=100) compared with the S&P 100 (about
365% APR for TI=90).
We
also note that the number of trades is dramatically
reduced as TI is increased. In the case
of the S&P 100, we only have 5 trades
for a TI threshold of 90 (33 for the S&P
500). What this means is that the when TI
is increased to 100, the S&P 100 profile
only shows 5 trades over the entire 4 year
period, and the S&P 500 shows 33 in
the same period.
Next
Research Steps for the Nirvana Club
Some would argue that even with 5 trades,
the fact that we are seeing 365% APR over
the test period is a superb result, and
getting 33 trades on the S&P 500 for
a 105% APR is also very good. In our opinion,
however, this is not a sufficiently large
"trade count" to make the profile
ready for automatic trading. We would prefer
to see something like 3-6 trades per stock
per year, in order to have enough candidates
to allow diversification. Current research
underway in the Nirvana Labs is focused
on maintaining high Tradability while increasing
trade count.
Once
trade count has been increased with the
same or higher profitability, performance
approaching Ultimate Trading Machine status
will be realized, since the system could
then be traded automatically to achieve
a strong, positive return with lessened
risk through diversification. That is the
goal of our next phase of research.
back
to top
Continue
to
"The Nirvana Club: A Unique Opportunity
for OmniTrader Users"
|
;)
;)
