Posted by dickdos2 on 15th of Sep 2011 at 03:55 pm
Perthx - I ran some numbers on this earlier today, looking at
all entries. For simplicity, let's say I'm looking at the
multi-entry system.
So far, for live trading, about half of the signals do not
execute. On average, the trigger go over/under by about 0.60% of
the prevailing SPY price. The misses tend to come in clusters
interestingly enough.
Like I said, I looked at total entries from the perspective
of the multi-entry system. The single entry system has a different
perspective.
Posted by dickdos2 on 15th of Sep 2011 at 11:17 am
Since the system was opened to new members in May (live trading
with signals sent to members), there have been a total of 26 entry
signals. This includes all entry signals, both single-entry and
multi-entry systems. Not all of these signals have been executed,
of course.
Of these 26 signals, 5 have been issued with a high-low range.
This works out to ~20% of all entry signals, which is certainly not
rare. But is this the norm you say? Well, if we assume that the
rate of potential signals generated over the life of the system is
mostly constant, then we can make some educated guesses about what
to expect going forward.
Of the 26 signals, 12 were executed. This is roughly two signals
for every actual entry. There have been 405 entries in the
multi-entry system to date. If we use some statistical analysis, we
see that 26 signals out of a population of 810 signals (assumed)
gives us at +/- 15% margin of error (at 95% confidence level).
This means: if our assumptions hold, then customers of the SPY
system should expect that signals will contain a range between 5%
and 35% of the time, given current data. This is a wide range, but
it tells us that we can reasonable expect ranged signals to be part
of the system, and that we should prepare accordingly. Over time
we'll have more live-trading data and be able to refine these
expectations.
TL;DNR - be mentally prepared for ranged signals to be a
part of the system, and plan accordingly.
I invite anyone to examine the data of course,
hereis a Google Docs spreadsheet.
I plugged this
"30 minute MACD 25,170,25"concept into NinjaTrader
this morning, just to see what would happen. I used the MACD
histogram as described here, taking trades upon a crossover of the
zero line with one confirming bar.
It's pretty straightforward: when trading SPY, taking both long
and short positions accordingly, the profit factor was 1.38 with
cumulative profit of 22.3% (frictionless) for the period beginning
6/1/2010 and ending today, 9/2/2011.
I compared the results to a daily chart of SPY over the same
time period. Comparing trades dates with the daily charts, I could
see that the signal worked well in periods of high volatility
(i.e., choppy markets). The signal resulted in whipsaws whenever it
encountered a strongly trending market.
Newsletter
Subscribe to our email list for regular free market updates
as well as a chance to get coupons!
The community is delayed by three days for non registered users.
Some stats...
just curious but...
Posted by dickdos2 on 15th of Sep 2011 at 03:55 pm
Perthx - I ran some numbers on this earlier today, looking at all entries. For simplicity, let's say I'm looking at the multi-entry system.
So far, for live trading, about half of the signals do not execute. On average, the trigger go over/under by about 0.60% of the prevailing SPY price. The misses tend to come in clusters interestingly enough.
Like I said, I looked at total entries from the perspective of the multi-entry system. The single entry system has a different perspective.
Since system launch, small windows have been common
I am not questioning, not complaining, not second guessing -- ...
Posted by dickdos2 on 15th of Sep 2011 at 11:17 am
Since the system was opened to new members in May (live trading with signals sent to members), there have been a total of 26 entry signals. This includes all entry signals, both single-entry and multi-entry systems. Not all of these signals have been executed, of course.
Of these 26 signals, 5 have been issued with a high-low range. This works out to ~20% of all entry signals, which is certainly not rare. But is this the norm you say? Well, if we assume that the rate of potential signals generated over the life of the system is mostly constant, then we can make some educated guesses about what to expect going forward.
Of the 26 signals, 12 were executed. This is roughly two signals for every actual entry. There have been 405 entries in the multi-entry system to date. If we use some statistical analysis, we see that 26 signals out of a population of 810 signals (assumed) gives us at +/- 15% margin of error (at 95% confidence level).
This means: if our assumptions hold, then customers of the SPY system should expect that signals will contain a range between 5% and 35% of the time, given current data. This is a wide range, but it tells us that we can reasonable expect ranged signals to be part of the system, and that we should prepare accordingly. Over time we'll have more live-trading data and be able to refine these expectations.
TL;DNR - be mentally prepared for ranged signals to be a part of the system, and plan accordingly.
I invite anyone to examine the data of course, hereis a Google Docs spreadsheet.
Volatility is a must
email to me from a member
Posted by dickdos2 on 2nd of Sep 2011 at 01:02 pm
I plugged this "30 minute MACD 25,170,25"concept into NinjaTrader this morning, just to see what would happen. I used the MACD histogram as described here, taking trades upon a crossover of the zero line with one confirming bar.
It's pretty straightforward: when trading SPY, taking both long and short positions accordingly, the profit factor was 1.38 with cumulative profit of 22.3% (frictionless) for the period beginning 6/1/2010 and ending today, 9/2/2011.
I compared the results to a daily chart of SPY over the same time period. Comparing trades dates with the daily charts, I could see that the signal worked well in periods of high volatility (i.e., choppy markets). The signal resulted in whipsaws whenever it encountered a strongly trending market.