It seems to me that adding stops to the system could lock in
large losses which don't exist in backtesting the straight sell
signals. I am not inclined to do anything that could change already
stellar backtest results.
If this is a reversion to the mean logic system and the first
buy is on a "stretched rubber band", the second is stretched a
little more, the third......, etc, then why sell just before
the snapback that gives our profits? Matt's system has a "stop" if
you want to call it that. IT's called a sell signal, no? Why would
I want to sell somewhere of my choosing unless I didn't trust that
my performance could come close to the backtest results?
Any modification is potentially dangerous to backtest results.
With such fantastic backtest results, I would first like to achieve
something similar (which would be a big improvement on my prior
results) before trying to make it better on my own !!
For me, it's just keep it simple and work the plan and establish
some consistent results and try not to complicate the situation.
Period.
You are right, we should trust the system in every
case. The March '08 entry would have been a nail biter for
a couple of days after the 4th entry. The trade turned out to
be a very good one and fact is that a huge loss has not
happend in more than a decade, but could happen anytime in the
future.
Stops
entry weighting
Posted by dougmil on 7th of May 2011 at 07:46 pm
It seems to me that adding stops to the system could lock in large losses which don't exist in backtesting the straight sell signals. I am not inclined to do anything that could change already stellar backtest results.
If this is a reversion to the mean logic system and the first buy is on a "stretched rubber band", the second is stretched a little more, the third......, etc, then why sell just before the snapback that gives our profits? Matt's system has a "stop" if you want to call it that. IT's called a sell signal, no? Why would I want to sell somewhere of my choosing unless I didn't trust that my performance could come close to the backtest results?
Any modification is potentially dangerous to backtest results. With such fantastic backtest results, I would first like to achieve something similar (which would be a big improvement on my prior results) before trying to make it better on my own !!
For me, it's just keep it simple and work the plan and establish some consistent results and try not to complicate the situation. Period.
Doug Milliken
stops...check this article out
Posted by perthx on 8th of May 2011 at 10:37 pm
maybe somebody at MarketWatch is reading the boards!! lol
ETF investors at risk in next ‘flash crash’
Commentary: Stop orders can’t stop losses in fast-moving markets
Article here
You are right, we should
Posted by zwyss on 8th of May 2011 at 02:55 am
You are right, we should trust the system in every case. The March '08 entry would have been a nail biter for a couple of days after the 4th entry. The trade turned out to be a very good one and fact is that a huge loss has not happend in more than a decade, but could happen anytime in the future.
Well said
Posted by druzin on 7th of May 2011 at 11:58 pm
Well said