Dont fool yourself by testing your algo over a long period, and
thinking long == good test.
the market probably trended up or down over your entire test.
if didnt, should it have? as the market has trend one way (if you ignore what most people call crashes but are only blips) for a long time...
How your algo will do in the next crash, needs testing on previous crashes.
(perhaps more than one, I don't know)
Another thing to try is that algo against others.
Assuming your algo held an average net postive positions (capital at risk)
In a particular how good did:
buys spread out over the first third and sells over the last third of the period do?
and various other 'dumb' control test algos.
controlling then for other algos that held for the same capital * length of time,
how did you do?
If your algo only, invests rarely and wisely, what is the opportunity cost of having capital available but not in play?
if not what What is the risk cost, of it being in play when the next crash comes...
The stats on shares, are nasty as there is lots of auto-correlation at long time scales.
be careful DYOR carefully, IMO be especially carefully when testing algo's.
ANZ Price at posting:
$34.81 Sentiment: Buy Disclosure: Held
A personalised tool to help users track selected stocks. Delivering real-time notifications on price updates, announcements, and performance stats on each to help make informed investment decisions.
(20min delay)