Trading Crude Light 5min Bars Using

The Adaptive Goertzel DFT System

Working Paper January 2014
Copyright  2014 Dennis Meyers

In the working paper entitled “Trading Crude Light 5min Bars Using The End Point Fast Fourier Transform ”,
http://www.meyersanalytics.com/publications2/CL5EPFFTXn.pdf , we demonstrated how to use the EPFFT
technique on a walk forward basis to trade 5-minute bars of Crude Light (CL) futures. Here it will be shown how to
use the Goertzel DFT(GZ) technique on a walk forward basis to trade 5-minute bars of Crude Light (CL) futures.

In a previous working paper entitled “MESA vs Goertzel DFT”,

http://www.meyersanalytics.com/publications2/MesaVsGDFT.pdf we demonstrated that the Goertzel Algorithm, a
subset of the Discrete Fourier Transform (DFT), has better frequency detection abilities of sine waves imbedded in
noise than the MESA algorithm when the noise amplitude is equal to or greater than the signal amplitude. MESA,
which stands for Maximum Entropy Spectral Analysis, is more commonly referred to as the Burg AR algorithm in
engineering and is a widely used mathematical technique designed to find the frequencies in data. MESA was
developed by J.P Burg for his Ph.D dissertation at Stanford University in 1975. The use of the MESA technique for
stocks and futures has been written about in many articles.

Previous researchers using MESA constrained themselves to using MESA to find only the cycle with the highest
amplitude and called that cycle the dominant cycle. In this paper we will find and use the frequencies with the five
or ten highest amplitudes to create a noise filtered signal curve that we will follow to create our strategy buy and sell
signals.

The Goertzel Algorithm.

The value of the Discrete Fourier Transform is that for N input data points, the DFT can not only find the
frequencies in the data but also the amplitude and phases of the sine waves in the data at the discrete frequency
points of 1/N, 2/N,…. to (N/2)/N. For instance if N, the number of price bars or closes, were equal to 20 we could
find the amplitude and phases for the 20 period cycle (20 bars/ cycle), the 10 period cycle, the 6 2/3 period cycle all
the way down to the 2 period cycle. We could then find the ten cycles with the highest amplitudes and reconstruct a
new signal with those ten periods, amplitudes and phases. Remember, cycle= 1/period. As traders we are more
accustomed to thinking of terms of periods like the 3 day cycle (frequency = 1/3 cycle/day) or the 10 bar cycle
(frequency = 1/10 cycles/bar) etc. Unfortunately the DFT can only calculate equally spaced frequencies of 1/N. In
the above example, where N=20, using the DFT we could only calculate the amplitudes of the periods
(periods=1/frequency) of 20, 10, 6 2/3 etc. What if the true signal period was between 20 and 10? The DFT
couldn’t find it. That was the advantage of MESA. MESA was not constrained to the 1/N spacing. With MESA any
grid of frequencies could be examined. Fortunately, as shown in my previous article, a special subset of the discrete
Fourier transform called the Goertzel algorithm can be used to find frequencies in between the 1/N frequency
divisions. The Goertzel algorithm is used extensively in tone detection. We all are familiar with cellular phones.
When you press a button have you ever wondered how the telephone company knows what button you pushed? The
answer is the Goertzel algorithm. This algorithm is built into tiny integrated circuits and immediately (within 43ms)
detects the tone of the button you pushed. Here we will use the Goertzel algorithm to detect the frequencies of the
price series we give it.

Despite the advantage of the Goertzel algorithm frequency detection abilities, it has three drawbacks. One drawback
of the Goertzel algorithm is that it is much slower than the Fast Fourier transform. If we had 512 data points and we
wanted to look at 128 different frequencies, the FFT computation would be proportional to 512*log2(512)= 4608
operations while the Goertzel computation would be proportional 512*128 = 65536 operations. In other words, in
this case, the Goertzel algorithm would take more than 10 times longer than the FFT to compute. However, as a
computation time comparison, Goertzel would take about half the time as MESA to compute those 128 frequencies.
The second drawback of the Goertzel algorithm is that in order to find a frequency in Goertzel when the noise
amplitude is high you need enough data so that your lowest frequency (largest period) is able to complete at least 3

Copyright © 2014 Dennis Meyers Trading Crude Light 5min Bars Using The Adaptive Goertzel DFT System Page 1 of 16
cycles. This means that if we were examining daily closing prices and we wanted to find periods of 75 days and less
than we would need at least 3*75=225 days of prices in order to detect that period in noisy data. The third drawback
of the Goertzel algorithm is that while it can detect the frequency within the 1/N spacing it cannot detect more than
one frequency within that spacing. For instance if N=20, and we are looking for a frequency between 1/20 and 2/20,
the Goertzel can detect the frequency anywhere at or between these two frequencies. However, if there were a
second frequency in-between these two frequencies, Goertzel could not find it. If there were two frequencies
between these two values, Goertzel would produce one frequency as a weighted average of the two. This is where
MESA has a clear advantage. On data with a low noise component, Mesa could detect these two closely spaced
frequencies. However, if the data is very noisy than MESA would not be able to detect the closely spaced
frequencies either.

The Adaptive n Cycle Goertzel-DFT System.

The nature of intraday price movements are constantly changing due to current economic surprises, events and trader
sentiment. Also the time of year changes the nature of intraday markets, such as the seasons, holidays, vacation time,
etc. As such, the periods or frequencies found on intraday prices 3 weeks ago may no longer be the same as the
frequencies found on today’s intraday data. We would expect the frequencies found on intraday data to vary over
time.

For this system we will create an indicator that walks forward one bar at a time. The indicator will take a fixed
number, N, of closing prices and use the Goertzel Algorithm(GA) to find the N frequencies with the highest
amplitudes. Using those frequency’s, amplitudes and phases, we will construct a new price that forecasts the price
one bar ahead. We will save this next bar forecast value, or forecast point. Next we will move the N closing prices
forward one bar adding the next bar and dropping the last bar so that we have exactly N closing prices again. With
this new N closing prices window we compute the next bar forecast value and save it. We keep marching the N
closing price window forward 1 bar at a time, calculate and save the new forecast point until we reach the end of our
data. We will then connect all the generated forecast points to produce a curve that creates the n cycle next bar
forecast as the ten frequencies used to create the next bar forecast change over time. Thus this curve adapts to the
closing price’s changing frequencies and projects one day ahead so that it’s lag is minimized when things change.

For example suppose we had daily data from 7/15/13cto 7/20/13. We would calculate and save the next day forecast
value of 7/20/13 and slide our data window up one day from 7/16/13 to 7/21/13. We would calculate the next day’s
forecast value of this new data window and save the forecast value of 7/21/13. We would keep sliding our data
window forward one day at a time, calculate the next day’s forecast value and save that forecast point. When we
reached the end of our data, we would then connect all these saved forecast points to create a new curve which we
would follow and use to create the system buy and sell signals.

Adaptive Goertzel DFT System Construction Details

Unfortunately constructing the n cycle DFT of a price data series is not quite as simple as just taking N closing
prices, and directly plugging them into a Goertzel algorithm.

The DFT assumes the time domain sample is periodic and repeats. Suppose a price series starts at 400 and wiggles
and wags for 512 data samples ending at the value of 600. The DFT assumes that the price series starts at zero,
suddenly jumps to 400, goes to 600 and suddenly jumps down to zero again and then repeats. The DFT must create
all kinds of different frequencies in the frequency domain to try and match this type of behavior. These false
frequencies created to match the jumps and the high average price completely swamp the amplitudes of any real
frequencies making them look like noise. Fortunately this effect can be almost eliminated by a simple technique
called end point flattening.

The calculation of end point flattening coefficients is simple. If x(1) represents the first price in the sampled data
series, x(n) represent the last point in the data series and y(i) equal to the new endpoint flattened series then:

a = x(1) b=(x(n)-x(1))/(n-1)

y(i) = x(i) – [a+ b*(i-1)] for i=1 to n (1)

Copyright © 2014 Dennis Meyers Trading Crude Light 5min Bars Using The Adaptive Goertzel DFT System Page 2 of 16
We can see that when i=1 then y(1)=0 and when i=n then y(n) =0. What we’ve done is subtract the beginning value
of the time series to make the first value equal to zero and then rotate the rest of the time series such that the end
point is now zero. This technique reduces the endpoint distortion but introduces a low frequency artifact into the
Fourier Frequency spectrum. Fortunately we won’t be looking for frequencies in that range so this distortion will
have minimal impact.

10 Cycle Goertzel-DFT Curve Construction

Before we start, we have to determine what is the largest period we will be looking to include in the 10 cycle
Goertzel DFT construction. For intraday data a 1-day cycle is very important. Higher periods are also important but
they don’t impact on intraday trading as much. For this article we will use 5-minute bars of the Crude Light(CL)
Futures. The CL trades on the CME Globex from 6pm EST to 5:15pm EST on the next day with a 45min break six
days a week from Sunday at 6pm EST to Friday at 5:15 EST. Each day consists of 279 5min bars a day. Thus the 1-
day cycle would need 279 bars. In order to detect the 1-day cycle we need the cycle to repeat at least three times.
This means we need at least 3*279 = 837 data points in order to detect the 1-day cycle using 5-minute bars. We will
use 837 five min bars of CL data.

For this article a sliding time-bar window of 837 five-minute bars of the CL futures from 12/01/11 to 1/03/14 will be
used.

We will use the Goertzel algorithm described in our previous article to determine the amplitude of the frequencies.
We will also use the Goertzel algorithm to determine the phases of the frequencies with the 10 highest amplitudes.

Step 1 End flatten the 837 prices using equation (1) above.

Step 2 Use the Goertzel algorithm to calculate the amplitudes for the frequencies of 1/279 down to 1/6. Frequency
= 1/period. We are scanning for periods because as traders we think in terms of periods not frequencies. That is, the
1 day cycle, the 20 bar cycle etc. Thus, here we are scanning for frequencies of 279 bars/cycle to 6 bars/cycle. The
frequencies of 5bars/cycle down to 2 bars/cycle move to fast with 5min bars to take advantage trading these cycles.
The slippage and commissions would eat up any profits made.

Step 3 Find the ten frequencies with the highest amplitudes, calculate the phases of these frequencies and save these
amplitudes (a[i]), phases (phi[i]) and frequencies (f[i]). Where [i] is one of the ten highest amplitudes found.

Step 4 Calculate the forecast next bar value and the end point bar value using the above ten frequencies, amplitudes
and phases. The forecast(fp) = ∑a[i]*cos(2*PI*f[i]*838 +phi[i]) i=1 to 10 and the Endpoint(ep) =
∑a[i]*cos(2*PI*f[i]*837 +phi[i]) i=1 to 10 where a[i], f[i] and phi[i] are the ten frequency, amplitude and phases
found.

Step 5 Save the calculated forecast next bar point and the end point values. Call the forecast next point fp(k) and the
end point ep(k) where k denotes the order of the sliding window. That is, the first sliding window k=1, the second,
k=2, etc. Slide the 837 bar data window forward one bar, and repeat steps 1 through 4.

Why do we need fp and ep? When the data window is moved forward one bar at a time a new data sample is added
to the end and the data sample at the beginning is subtracted. This adding and subtracting causes the end point
flattening coefficients and the power in the frequency spectrum to jump around creating distortion and jitter in the
calculation of the forecast next bar point. This random jumping as the data window slides forward in time adds a
small random jump to the forecast next bar point curve. Fortunately this jumping can be minimized by creating a
curve from the two saved end points, fp(k) and ep(k), above in step 5. Since turning points are of interest rather
than magnitude then in step 5 a new variable will be created called sumv where

sumv(k) = sumv(k-1) + fp(k)-ep(k)

Copyright © 2014 Dennis Meyers Trading Crude Light 5min Bars Using The Adaptive Goertzel DFT System Page 3 of 16
fp(k)-ep(k) is like a one bar ahead momentum or velocity. This new curve sumv(k) is the sum of all the changes in
the next bar’s ten cycle forecast value fp(k) from the end point ten cycle value. This change series minimizes the
magnitude jump problem creating a fairly smooth momentum sum curve.

The Adaptive 10 Cycle Goertzel-DFT System Defined

Even though sumv is a fairly smooth curve it still has a number of short term wiggles preventing us from simply
going long when the curve turns up and going short when the curve turns down. To create a system, we will use a
simple curve following technique on CL 5-minute bars.

Buy Rule:
• IF sumv has moved up by more than the point amount of pntup from the lowest low recorded in sumv while
short then buy the CL futures at the market..

Sell Rule:
• IF sumv has moved down by more than the point amount pntdn from the highest high recorded in sumv while
long then sell the CL futures at the market.

Intraday Bars Exit Rule:

Close the position at 1420 ET before the CL open outcry close at 1430 ET (no trades will be carried overnight).

Intraday Bars First Trade of Day Entry Rule:

Ignore all trade signals before 8:30am ET . For the Buy and Sell rules above we have included a first trade of the day
entry rule. Trading in the CL futures has changed a lot in the last 5 years because of 24hr Globex trading. In
particular trading starts a lot earlier in the morning when Asia and then Europe opens and then dies down. The CL
volume starts to pick up again around 830am ET before the open outcry floor opens at 9am ET so we will have the
strategy resume trading at 830am ET

Discussion of Crude Light Future Prices

Crude Light futures (CL) are traded on Globex and on the trading floor at the NYMEX. On Globex the CL is traded
on a 23hour basis . The NYMEX hours for open outcry floor trading (RTH) are 9:00 ET to 14:30 ET. Over 60% of
the volume in the CL is done on Globex during the NYMEX open outcry hours. We have restricted our study to
only trading the CL during the 8:30 to 14:20 ET hours.

Testing The Adaptive N Cycle Goertzel-DFT (GZ) System Using Walk Forward Optimization
There will be three strategy parameters to determine:
1. ncy, Number of Cycles (for this study we will look at both 10 and 5 cycles).
2. pntup, if sumv has moved up by more than the point amount of pntup from the lowest low recorded in
sumv while short then issue a buy signal
3. pntdn, if sumv has moved down by more than the point amount pntdn from the highest high recorded in
sumv while long then sell

To test this system we will use five minute bar prices of Crude Light (CL) futures contract traded on the
NYMEX/Globex and known by the symbol CL for the 109 weeks from December 1, 2011 to January 3, 2014.

We will test this strategy with the above CL 5 min bars on a walk forward basis, as will be described below. To
create our walk forward files we will use the add-in software product called the Power Walk Forward Optimizer
(PWFO). In TradeStation (TS), we will run the PWFO strategy add-in along with GZ Strategy on the CL 5min bar
data from December 1, 2011 to January 3, 2014..

What Is A In-sample section and Out-Of-Sample Section?

Whenever we do a TS optimization on a number of different strategy inputs, TS generates a report of performance
metrics (total net profits, number of losing trades, etc) vs these different inputs. If the report is sorted on say the total
net profits(tnp) performance metric column then the highest tnp would correspond to a certain set of inputs. This is
called a in-sample section. If we choose a set of strategy inputs from this report based upon some performance

Copyright © 2014 Dennis Meyers Trading Crude Light 5min Bars Using The Adaptive Goertzel DFT System Page 4 of 16
metric we have no idea whether these strategy inputs will produce the same results on future price data or data they
have not been tested on. Price data that is not in the in-sample section is defined as out-of-sample data. Since the
performance metrics generated in the in-sample section are mostly due to “curve fitting” (see Walk Forward Out-of-
Sample Testing section below) it is important to see how the strategy inputs chosen from the in-sample section
perform on out-of-sample data.

What Does The Power Walk Forward Optimizer (PWFO) Do?

The PWFO is a TS add-in that breaks up the TS optimization run into a number of user selectable test and out-of-
sample sections. The PWFO prints out the test sample performance and the out-of-sample performance results,
on one line, for each case or input variable combination that is run by the TradeStation(TS) optimization module to a
user selected spreadsheet comma delimited file. The PWFO can generate up to 500 different test and out-of-sample
date optimization files in one TS run, saving the user from having to generate optimization runs one at a time. The
PWFO output allows you to quickly determine whether your procedure for selecting input parameters for your
strategy just curve fits the price and noise, or produces statistically valid out-of-sample results. In addition to the
out-of-sample performance results presented for each case, 30+ superior and robust performance metrics (many are
new and never presented before) are added to each case line in the in-sample section and printed out to the comma
delimited file. These 30+ performance metrics allow for a superior and robust selection of input variables from the
in-sample section that have a higher probability of performing well on out-of-sample data (Please see Appendix 2
for a listing of these performance metrics).

For our computer run we will have the PWFO breakup the 109 weeks of CL five-minute bar price data into 105
test/out-of sample files. The in-sample sections will be 30 calendar days and the out-of-sample(oos) section will be
the one week following the in-sample section. The oos week will always end on a Friday as will the 30-day calendar
in-sample section. As an example the first in-sample section would be from 12/1/2011 to 12/30/2011 and the out-
of-sample section would be from 1/2/2012 to 1/6/2012.(our test and out-of-sample sections always end on a Friday).
We would then move everything ahead a week and the 2nd in-sample section would be from 12/8/2011 to 1/6/2012
and the out-of-sample section would be from 1/9/2012 to 1/13/2012. Etc.

The PWFO 105 test/out-of-sample section dates are shown in Table 1 on page 8 below. We will then use another
software product called the Walk Forward Performance Metric Explorer (WFME) on each of the 105 test and out-
of-sample(oos) sections generated by the PWFO to find the best in-sample section performance filter that
determines the system input parameters (ncy, pntup, pntdn) that will be used on the out-of-sample data. Detailed
information about the PWFO and the WFME can be found at www.meyersanalytics.com

For the test data we will run the TradeStation combinatorial optimization engine on the 105 weeks of CL 5 min bars
with the following ranges for the Goertzel(GZ) strategy input variables.
1. ncy 5 and 10
2. pntup from 0.2 to 3 steps of 0.2
3. pntdn from 0.2 to 3 in steps of 0.2
4. myStartTime = 830
5. Xtime = 1420

This will produce 450 different cases or combinations of the input parameters for each of the 105 PWFO output
files.

Walk Forward Out-of-Sample Testing

Walk forward analysis attempts to minimize the curve fitting of price noise by using the law of averages from the
Central Limit Theorem on the out-of-sample performance. In walk forward analysis the data is broken up into many
test and out-of-sample sections. Usually for any system, one has some performance metric selection procedure,
which we will call a filter, used to select the input parameters from the optimization run. For instance, a filter might
be all cases that have a profit factor (PF) greater than 1 and less than 3. For the number of cases left, we might select
the cases that had the best percent profit. This procedure would leave you with one case in the in-sample section
output and it's associated strategy input parameters. Now suppose we ran our optimization on each of our many in-
sample sections and applied our filter to each in-sample section output. We would then use the strategy input

Copyright © 2014 Dennis Meyers Trading Crude Light 5min Bars Using The Adaptive Goertzel DFT System Page 5 of 16
parameters found by the filter in each in-sample section on the out-of-sample section immediately following that in-
sample section. The input parameters found in each in-sample section and applied to each out-of-sample section
would produce independent net profits or losses for each of the out-of-sample sections. Using this method we now
have "x" number of independent out-of-sample section profit and losses from our filter. If we take the average of
these out-of-sample section net profits and losses, then we will have an estimate of how our system will perform on
average. Due to the Central Limit Theorem, as our sample size increases, the spurious noise results in the out-of-
sample section performance tend to average out to zero in the limit leaving us with what to expect from our system
and filter. Mathematical note: This assumption assumes that the out-of-sample returns are from probability
distributions that have a finite variance.

Why use the walk forward technique? Why not just perform an optimization on the whole price series and choose the
input parameters that give the best total net profits or profit factor? Surely the price noise cancels itself out with
such a large number of test prices and trades. Unfortunately, nothing could be farther from the truth! Optimization
is a misnomer and should really be called combinatorial search. As stated above, whenever we run a combinatorial
search over many different combinations of input parameters on noisy data on a fixed number of prices, no matter
how many, the best performance parameters found are guaranteed to be due to “curve fitting” the noise and signal.
What do we mean by “curve fitting”? The price series that we trade consists of random spurious price movements,
which we call noise, and repeatable price patterns (if they exist). When we run, for example, 5000 different inputs
parameter combinations , the best performance parameters will be from those system input variables that are able to
produce profits from the price pattern and the random spurious movements While the price patterns will repeat, the
same spurious price movements will not. If the spurious movements that were captured by a certain set of input
parameters were a large part of the total net profits, then choosing these input parameters will produce losses when
traded on future data. These losses occur because the spurious movements will not be repeated in the same way.
This is why system optimization or combinatorial searches with no out-of-sample testing cause loses when traded in
real time from something that looked great in the in-sample section. Unfortunately it is human nature to extrapolate
past performance to project future trading results and thus results from curve fitting give the illusion, a modern “siren
call” so to speak, of future trading profits.

In order to gain confidence that our input parameter selection method using the optimization output of the test data
will produce profits, we must test the input parameters we found in the in-sample section on out-of-sample data. In
addition, we must perform the test/out-of-sample analysis many times. Why not just do the out-of-sample analysis
once? Well just as in Poker or any card game, where there is considerable variation in luck from hand to hand, walk
forward out-of-sample analysis give considerable variation in week-to-week out-of-sample profit “luck”. That is, by
pure chance we may have chosen some input parameter set that did well in the in-sample section data and the out-of-
sample section data. In order to minimize this type of “luck”, statistically we must repeat the walk forward out-of-
sample (oos) analysis over many test/oos sections and take an average of our weekly results over all out-of-sample
sections. This average gives us an expected weekly return and a standard deviation of weekly returns which allows
us to statistically estimate the expected equity and it’s range for N weeks in the future.

Finding The Strategy Input Parameters in The Walk Forward In-sample sections
The PWFO generates a number of performance metrics in the in-sample section. The question we are attempting to
answer statistically, is which performance metric or combination of performance metrics (which we will call a filter)
in the in-sample section will produce strategy inputs that produce statistically valid profits in the out-of-sample
section. In other words we wish to find a metric filter that we can apply to the in-sample section that can give us
strategy inputs that will produce, on average, good trading results in the future. The PWFO produces a total of 32
different performance metrics in the in-sample section. If we have 450 different input variations or cases then the
in-sample section consists of 32 columns of performance metrics for each of the 450 input cases or rows.

An example of a simple filter would be to choose the row in the in-sample section that had the highest net profit or
perhaps a row that had one the best performance metric from one of the other 32 PWFO metrics. Unfortunately it
was found that this type of simple filter very rarely produces good out-of-sample results. More complicated metric
filters can produce good out-of-sample results minimizing spurious price movement biases in the selection of
strategy inputs.

Copyright © 2014 Dennis Meyers Trading Crude Light 5min Bars Using The Adaptive Goertzel DFT System Page 6 of 16
Here is an example of a better more complicated filter that was used in this here. Not many traders can stay with a
strategy that has a large number of losers in a row (LR). For this filter we will choose LR<=3. This choice of LR is
completely arbitrary and is what I feel comfortable with. After using the LR filter, as described, there can still be
100’s of rows left in the PWFO file in-sample section. The PWFO generates the metric mWb. Wb is the number of
bars in each of the winning trades. This metric calculates the median of all the Wb. We use the median for this
metric, because we do not want the statistic distorted by an outlier trades. Thus we would want the median to be as
large as possible. Let us choose the 5 rows that contain the largest(top) mWb values from the rows that are left
from the LR screen. In other words we sort mWb from high to low, eliminate the rows that have LR>3 then choose
the top 5 Rows of whatever is left. This particular filter will now leave 5 cases or rows in the PWFO file that satisfy
the above filter conditions. We call this filter t5mWb|lr3 where t5mWb means the Top 5 mWb rows left after the
LR filter.. Suppose for this filter, within the 5 PWFO rows that are left, we want the row that has the minimum
PWFO mLTr metric in the in-sample section. The metric mLTr is the median of all the losing trades in the in-
sample section. Again we use the median rather than the average because we do not want this statistic distorted by
outlier trades. This would produce a filter named t5mWb|lr3-mLTr. This filter leaves only one row in the PWFO
in-sample section with its associated strategy inputs and out-of-sample net profit in the out-of-sample section. This
particular t5mWb|lr3-mLTr filter is then calculated for each of the 105 PWFO files and the average 105 out-of-
sample performance is calculated. In addition many other important out-of-sample performance statistics for this
filter are calculated and summarized. Figure 3 shows such a computer run along with a small sample of other filter
combinations that are constructed in a similar manner. Row 3 of the sample output in Figure 3 shows the results of
the filter discussed above.

Bootstrap Probability of Filter Results. Using modern "Bootstrap" techniques, we can calculate the probability of
obtaining each filter's total out-of-sample net profits by chance. By net we mean subtracting the cost and slippage of
all round trip trades from the total out-of-sample profits. Here is how the bootstrap technique is applied. Suppose as
an example, we calculate the total out-of-sample net profits(tOnpNet) of 5000 different TopN-Metric-LR filters. A
mirror filter is created for each of the 5000 filters. However, instead of picking an out-of-sample net profit(OSNP)
from a filter row as above, the mirror filter picks a random row's OSNP in each of the 105 PWFO files. Each of the
5000 mirror filters will choose a random row's OSNP of their own in each of the 105 PWFO files.. At the end, each
of the 5000 mirror filters will have 105 random OSNP's picked from the rows of the 105 PWFO files. The sum of
the 105 random OSNP picks for each mirror filter will generate a random total out-of-sample net profit(tOnpNet).
The average and standard deviation of the 5000 mirror filter's different random tOnpNets will allow us to calculate
the chance probability of each TopN-Metric-LR filter's tOnpNet. Thus given the mirror filter's bootstrap random
tOnpNet average and standard deviation, we can calculate the probability of obtaining the TopN-Metric-LR- filter's
tOnpNet by pure chance alone. Since for this run there are 3844 different filters, we can calculate the expected
number of cases that we could obtain by pure chance that would match or exceed the tOnpNet of the filter we have
chosen or (3844) X (tOnpNet Probability). For our filter, in row 3 in Figure 3, the expected number of cases that we
could obtain by pure chance that would match or exceed the $59260 is 3844 x 6.13 10-5 = 0.236 . This is much less
than one case so it is improbable that our result was due to pure chance.

The partial run shown in Figure 3 reveals that the following filter will produce the most consistent and reliable out-
of-sample results.

Filter: LR<=3 and Top 5 mWb then minimum mLTr

Where:
• LR = Maximum losing trades in a row for a given set of inputs in the in-sample section
• mWb= Wb is the number of bars in each of the winning trades. This metric calculates the median of all the Wb
• mLTr = The metric mLTr is the median of all the losing trades in the in-sample section
The first part of the filter chooses those rows or cases out of the 450 rows in each PWFO file test(in-sample) section
that satisfy the criteria LR<=3. After using the LR filter, there can still be 100’s of rows left in the PWFO file. The
PWFO generates the metric mWb, which is the median of the number of bars in each of the winning trades. Let us

Copyright © 2014 Dennis Meyers Trading Crude Light 5min Bars Using The Adaptive Goertzel DFT System Page 7 of 16
choose the 5 rows that contain the Top 5 mWb values from the rows that are left from the LR screen.. This
particular filter will now leave 5 cases or rows in the PWFO file that satisfy these filter conditions. We call this filter
t5mWb|lr3 . Within the 5 PWFO rows that are left, we want the row that has the minimum PWFO metric, mLTr, in
the in-sample section. This Filter or selection procedure will leave only one choice for the startegy input values of
ncy, pntup, pntdn. We then use these input values found in the in-sample section by the Filter on the next week of
five minute CL out-of-sample price bars following the in-sample section.

Results
Table 1 below presents a table of the 105 test and out-of-sample windows , the selected optimum parameters and the
weekly out-of-sample profit/loss results using the filter described above.

Figure 1 presents a graph of the equity curve generated by using the filter on the 105 weeks of 1/06/12 – 1/03/14.
The equity curve is plotted from NetEq in Table 1. Plotted on the equity curve is the least squares straight line. The
weekly CL prices are also included on the plot so that we can compare how the equity curve changes vs CL price
changes.

Figure 2 presents the out-of-sample 5 minute bar chart of CL for 12/30/13 to 1/3/14 with the GZ Indicator and all
the buy and sell signals for those dates.

Figure 3 Partial output of the Walk Forward Metric Performance Explorer

Discussion of System Performance

In Figure 3 Row 2 of the spreadsheet filter output are some statistics that are of interest for our filter An interesting
statistic is Blw. Blw is the maximum number of weeks the OSNP equity curve failed to make a new high. Blw is 12
weeks for this filter. This means that 12 weeks was the longest time that the equity for this strategy failed to make a
new equity high.

To see the effect of walk forward analysis, take a look at Table 1. Notice how the input parameters ncy, pntup and
pntdn take sudden jumps from high to low and back . This is the walk forward process quickly adapting to
changing volatility conditions in the test sample. In addition, notice how often ncy changes from 5 to 10.

In Figure 1, which presents a graph of the equity curve using the filter on the 105 weeks of 1/06/12 – 1/03/14, notice
how the equity curve follows the trend line with an R2 of 0.96. The weekly plot of CL prices is superimposed on the
equity curve plot. Notice how fast the equity curve is able to adapt with changing trends in CL. From 2/24/12 to
6/22/12 CL fell fast from 110 to 80. From 6/22/12 to 9/14/12 CL climbed fast from 80 to 99. Yet despite these
changing price trends the GZ strategy was able to adapt and continue to generate net profits.

Using this filter, the strategy was able to generate $51050, net equity after commissions and slippage trading one CL
contract for 105 weeks. This period of time from 1/06/12 – 1/03/14 was at volatile period for crude oil. Yet the GZ
strategy was able to adapt quite well. Note $30 roundtrip commission and slippage was subtracted from each trade
and no positions were carried over night. From Table 1, the largest net losing week was -$3880 on 5/10/13 a very
whipsaw market week. The largest net equity drawdown was -$9970 during whipsaw CL price movements from
12/28/12-3/15/13. The longest time between new equity highs was 12 weeks.

In observing Table 1 we can see that this strategy and filter made trades from a low of no trades per week to a high
of 9 trades/week with an average of 4.5 trades/week. Note that the GZ strategy did not trade in 14 of the 105
weeks.

Given 24 hour trading of the Crude Light, restricting the strategy to trade only from 830am to 2:20pm ET caused the
strategy to miss many profitable trends opportunities when Asia and then Europe opened trading in the early
morning. As can be seen from the CL 5min bar chart in Figure 2, while the strategy indicated to remain short
restricting trading from 830am to 2:20pm missed the main overnight trend and the strategy profits were less then
they could have been if overnight trading was allowed. Further research will include the A.M. time zones.

Copyright © 2014 Dennis Meyers Trading Crude Light 5min Bars Using The Adaptive Goertzel DFT System Page 8 of 16
Disclaimer
The strategies, methods and indicators presented here are given for educational purposes only
and should not be construed as investment advice. Be aware that the profitable performance
presented here is based upon hypothetical trading with the benefit of hindsight and can in no way
be assumed nor can it be claimed that the strategy and methods presented here will be profitable
in the future or that they will not result in losses.

References:
1. Burg, J. P., ‘Maximum Entropy Spectral Analysis”, Ph.D. dissertation, Stanford University, Stanford, CA. May
1975.
2. Goertzel, G., “An Algorithm for he evaluation of finite trigonometric series” American Math Month, Vol 65,
1958 pp34-35.
3. Kay, Steven M., “Modern Spectral Estimation”, Prentice Hall, 1988
4. Marple, Lawrence S. Jr., “Digital Spectral Analysis With Applications”, Prentice Hall, 1987
5. Meyers, Dennis , “MESA vs Goertzel DFT”, Working Paper, http://meyersanalytics.com/articles.htm
6. Press, William H., et al, “Numerical Receipts in C++: the Art of Scientific Computing”, Cambridge Press, 2002.
7. Oppenheim, A, Schafer, R. and Buck, J., “Discrete Time Signal Processing”, Prentice Hall, 1996, pp663-634
8. Proakis, J. and Manolakis, D. “Digital Signal Processing-Principles, Algorithms and Applications”, Prentice
Hall, 1996., pp480-481

Info on Dennis Meyers

Dennis Meyers ([email protected].) has a doctorate in applied mathematics in engineering. He is a private
trader, and president of Meyers Analytics (www.MeyersAnalytics.com). His firm specializes in Financial
Engineering consulting for financial institutions and developing publicly available analytical software for traders.

Copyright © 2014 Dennis Meyers Trading Crude Light 5min Bars Using The Adaptive Goertzel DFT System Page 9 of 16
 Table 1 Walk Forward Out-Of-Sample Performance Summary for
CL End Point Fast Fourier Transform System

CL-5 min bars 11/30/2011 - 1/03/2014. The input values ncy, pntup, pntdn are the values found
from applying the filter to the in-sample section optimization runs.
In-Sample Dates Out-Of_Sample Dates osnp NOnp$30 NetEq ollt odd ont avosnp ncy pup pdn sTime xt CL CLS
11/30/11 to 12/30/11 01/02/12 to 01/06/12 130 10 10 -1280 -1850 4 32.5 5 1 2.4 830 1420 101.93
12/07/11 to 01/06/12 01/09/12 to 01/13/12 (3760) (3880) (3870) -2360 -3760 4 -940 10 1.2 2.4 830 1420 99.03
12/14/11 to 01/13/12 01/16/12 to 01/20/12 3470 3320 (550) -360 -360 5 694 5 2 1 830 1420 98.75
12/21/11 to 01/20/12 01/23/12 to 01/27/12 1230 1110 560 -70 -70 4 307.5 10 1.4 1.6 830 1420 100.12
12/28/11 to 01/27/12 01/30/12 to 02/03/12 (2650) (2740) (2180) -2700 -3780 3 -883.3 10 1.6 1.6 830 1420 98.13
01/04/12 to 02/03/12 02/06/12 to 02/10/12 (2180) 5 2.2 3 830 1420 99.41
01/11/12 to 02/10/12 02/13/12 to 02/17/12 (2180) 5 2.2 2.8 830 1420 104.42
01/18/12 to 02/17/12 02/20/12 to 02/24/12 360 240 (1940) -1040 -1040 4 90 10 1.2 1 830 1420 109.62
01/25/12 to 02/24/12 02/27/12 to 03/02/12 (1940) 5 2.2 2.8 830 1420 106.5
02/01/12 to 03/02/12 03/05/12 to 03/09/12 (1480) (1660) (3600) -1300 -1750 6 -246.7 5 0.8 1.4 830 1420 107.94
02/08/12 to 03/09/12 03/12/12 to 03/16/12 (3600) 5 1.8 3 830 1420 107.67
02/15/12 to 03/16/12 03/19/12 to 03/23/12 (420) (450) (4050) 0 0 1 -420 5 1.8 3 830 1420 106.75
02/22/12 to 03/23/12 03/26/12 to 03/30/12 (330) (420) (4470) -420 -420 3 -110 5 1 2.4 830 1420 102.93
02/29/12 to 03/30/12 04/02/12 to 04/06/12 2890 2800 (1670) 0 0 3 963.3 5 0.8 2.6 830 1420 103.25
03/07/12 to 04/06/12 04/09/12 to 04/13/12 (230) (380) (2050) -1320 -1320 5 -46 5 0.8 1.6 830 1420 102.84
03/14/12 to 04/13/12 04/16/12 to 04/20/12 1320 1170 (880) -360 -360 5 264 10 0.6 1.2 830 1420 104.04
03/21/12 to 04/20/12 04/23/12 to 04/27/12 840 690 (190) -270 -490 5 168 10 0.6 1.2 830 1420 104.81
03/28/12 to 04/27/12 04/30/12 to 05/04/12 2900 2720 2530 -1740 -2400 6 483.3 10 1.2 1.2 830 1420 98.59
04/04/12 to 05/04/12 05/07/12 to 05/11/12 (1030) (1300) 1230 -1100 -2490 9 -114.4 10 0.4 1 830 1420 95.57
04/11/12 to 05/11/12 05/14/12 to 05/18/12 1350 1230 2460 0 0 4 337.5 5 1.2 0.6 830 1420 91.65
04/18/12 to 05/18/12 05/21/12 to 05/25/12 1360 1240 3700 -200 -200 4 340 10 1.4 0.6 830 1420 90.72
04/25/12 to 05/25/12 05/28/12 to 06/01/12 3600 3450 7150 -200 -200 5 720 10 1.4 0.8 830 1420 83.26
05/02/12 to 06/01/12 06/04/12 to 06/08/12 800 590 7740 -730 -1400 7 114.3 5 1 0.6 830 1420 84.34
05/09/12 to 06/08/12 06/11/12 to 06/15/12 760 610 8350 -910 -1440 5 152 10 2 0.8 830 1420 84.84
05/16/12 to 06/15/12 06/18/12 to 06/22/12 1570 1390 9740 -1500 -1500 6 261.7 5 1 0.2 830 1420 80.15
05/23/12 to 06/22/12 06/25/12 to 06/29/12 6310 6160 15900 -380 -520 5 1262 5 1 0.6 830 1420 84.87
05/30/12 to 06/29/12 07/02/12 to 07/06/12 10 (140) 15760 -1160 -1840 5 2 10 0.6 2 830 1420 84.12
06/06/12 to 07/06/12 07/09/12 to 07/13/12 15760 5 2.8 2.6 830 1420 87.37
06/13/12 to 07/13/12 07/16/12 to 07/20/12 15760 5 2.8 2.6 830 1420 91.7
06/20/12 to 07/20/12 07/23/12 to 07/27/12 (1490) (1520) 14240 0 0 1 -1490 5 2.8 2.2 830 1420 90.26
06/27/12 to 07/27/12 07/30/12 to 08/03/12 1540 1480 15720 0 0 2 770 5 3 2.4 830 1420 91.39
07/04/12 to 08/03/12 08/06/12 to 08/10/12 (460) (520) 15200 -570 -570 2 -230 10 2 1.8 830 1420 93.7
07/11/12 to 08/10/12 08/13/12 to 08/17/12 (1370) (1400) 13800 0 0 1 -1370 5 1.4 2.4 830 1420 96.52
07/18/12 to 08/17/12 08/20/12 to 08/24/12 1790 1640 15440 -540 -540 5 358 5 1.4 2.4 830 1420 96.12
07/25/12 to 08/24/12 08/27/12 to 08/31/12 1950 1830 17270 -380 -620 4 487.5 5 1.4 2.4 830 1420 96.56
08/01/12 to 08/31/12 09/03/12 to 09/07/12 2270 2150 19420 -530 -530 4 567.5 5 1.6 2.4 830 1420 96.29
08/08/12 to 09/07/12 09/10/12 to 09/14/12 10 (140) 19280 -560 -560 5 2 10 2.2 2.8 830 1420 99.06
08/15/12 to 09/14/12 09/17/12 to 09/21/12 2030 1850 21130 -980 -1060 6 338.3 5 0.8 1 830 1420 93.39
08/22/12 to 09/21/12 09/24/12 to 09/28/12 (720) (900) 20230 -1300 -1300 6 -120 10 1.2 0.6 830 1420 92.53
08/29/12 to 09/28/12 10/01/12 to 10/05/12 (310) (460) 19770 -2530 -2880 5 -62 5 1.8 3 830 1420 90.35
09/05/12 to 10/05/12 10/08/12 to 10/12/12 2660 2540 22310 0 0 4 665 10 2.2 1 830 1420 92.07
09/12/12 to 10/12/12 10/15/12 to 10/19/12 1130 980 23290 -780 -780 5 226 10 2 0.8 830 1420 90.6
09/19/12 to 10/19/12 10/22/12 to 10/26/12 3980 3860 27150 0 0 4 995 10 2 0.8 830 1420 86.24
09/26/12 to 10/26/12 10/29/12 to 11/02/12 810 660 27810 -920 -1180 5 162 10 3 0.8 830 1420 84.79
10/03/12 to 11/02/12 11/05/12 to 11/09/12 1690 1600 29410 -1670 -1670 3 563.3 10 2.2 0.8 830 1420 86.17
10/10/12 to 11/09/12 11/12/12 to 11/16/12 (210) (360) 29050 -980 -1410 5 -42 10 2.8 0.8 830 1420 86.99
10/17/12 to 11/16/12 11/19/12 to 11/23/12 660 540 29590 -1090 -1090 4 165 10 2.2 1.8 830 1420 88.26

Copyright © 2014 Dennis Meyers Trading Crude Light 5min Bars Using The Adaptive Goertzel DFT System Page 10 of 16
 In-Sample Dates Out-Of_Sample Dates osnp NOnp$30 NetEq ollt odd ont avosnp ncy pup pdn sTime xt CL CLS
10/24/12 to 11/23/12 11/26/12 to 11/30/12 (70) (100) 29490 0 0 1 -70 10 2.2 2.2 830 1420 88.94
10/31/12 to 11/30/12 12/03/12 to 12/07/12 1710 1560 31050 -450 -590 5 342 10 2.2 2.2 830 1420 85.98
11/07/12 to 12/07/12 12/10/12 to 12/14/12 1430 1280 32330 -460 -460 5 286 10 2.2 2.2 830 1420 87.28
11/14/12 to 12/14/12 12/17/12 to 12/21/12 (110) (140) 32190 0 0 1 -110 5 2.2 2.6 830 1420 88.96
11/21/12 to 12/21/12 12/24/12 to 12/28/12 32190 5 1.4 2 830 1420 90.62
11/28/12 to 12/28/12 12/31/12 to 01/04/13 (30) (60) 32130 0 0 1 -30 5 1.4 2.2 830 1420 93.19
12/05/12 to 01/04/13 01/07/13 to 01/11/13 (380) (530) 31600 -560 -1110 5 -76 10 1.2 1.4 830 1420 93.77
12/12/12 to 01/11/13 01/14/13 to 01/18/13 31600 5 1.6 2.6 830 1420 95.79
12/19/12 to 01/18/13 01/21/13 to 01/25/13 680 530 32130 -850 -980 5 136 10 2 1.4 830 1420 96.09
12/26/12 to 01/25/13 01/28/13 to 02/01/13 (620) (710) 31420 -480 -620 3 -206.7 5 0.8 1.2 830 1420 97.61
01/02/13 to 02/01/13 02/04/13 to 02/08/13 (3010) (3160) 28260 -1640 -3160 5 -602 5 0.4 1 830 1420 95.76
01/09/13 to 02/08/13 02/11/13 to 02/15/13 120 (60) 28200 -1020 -1470 6 20 5 0.2 1.4 830 1420 97
01/16/13 to 02/15/13 02/18/13 to 02/22/13 (2630) (2780) 25420 -1900 -2670 5 -526 5 0.4 1.4 830 1420 93.73
01/23/13 to 02/22/13 02/25/13 to 03/01/13 (1880) (2030) 23390 -1250 -1880 5 -376 5 1 3 830 1420 91.39
01/30/13 to 03/01/13 03/04/13 to 03/08/13 (530) (710) 22680 -750 -1270 6 -88.3 10 0.8 0.2 830 1420 92.24
02/06/13 to 03/08/13 03/11/13 to 03/15/13 (310) (460) 22220 -750 -750 5 -62 5 0.8 0.2 830 1420 93.82
02/13/13 to 03/15/13 03/18/13 to 03/22/13 2300 2150 24370 -90 -90 5 460 10 0.6 1 830 1420 93.8
02/20/13 to 03/22/13 03/25/13 to 03/29/13 24370 5 3 3 830 1420 97.23
02/27/13 to 03/29/13 04/01/13 to 04/05/13 1440 1320 25690 -280 -480 4 360 5 0.8 1 830 1420 93.05
03/06/13 to 04/05/13 04/08/13 to 04/12/13 (70) (100) 25590 0 0 1 -70 5 1.2 1.6 830 1420 90.66
03/13/13 to 04/12/13 04/15/13 to 04/19/13 1310 1160 26750 -620 -620 5 262 5 1.8 2.8 830 1420 88.27
03/20/13 to 04/19/13 04/22/13 to 04/26/13 4210 4060 30810 -240 -240 5 842 5 1.8 2.8 830 1420 92.78
03/27/13 to 04/26/13 04/29/13 to 05/03/13 4050 3930 34740 -760 -760 4 1012.5 5 1.8 2.8 830 1420 95.61
04/03/13 to 05/03/13 05/06/13 to 05/10/13 (3130) (3280) 31460 -980 -3130 5 -626 5 3 1.8 830 1420 96.26
04/10/13 to 05/10/13 05/13/13 to 05/17/13 1560 1470 32930 0 0 3 520 10 1.6 3 830 1420 96.27
04/17/13 to 05/17/13 05/20/13 to 05/24/13 1930 1810 34740 -220 -220 4 482.5 10 1.8 3 830 1420 93.87
04/24/13 to 05/24/13 05/27/13 to 05/31/13 (1980) (2100) 32640 -1640 -2000 4 -495 5 1.8 2.6 830 1420 91.61
05/01/13 to 05/31/13 06/03/13 to 06/07/13 2080 1930 34570 -160 -160 5 416 5 0.2 1.4 830 1420 96.26
05/08/13 to 06/07/13 06/10/13 to 06/14/13 3000 2850 37420 0 0 5 600 10 0.8 1.4 830 1420 97.87
05/15/13 to 06/14/13 06/17/13 to 06/21/13 (3670) (3820) 33600 -1820 -3670 5 -734 10 2.4 1.6 830 1420 93.68
05/22/13 to 06/21/13 06/24/13 to 06/28/13 2550 2370 35970 -420 -420 6 425 10 0.4 0.8 830 1420 96.22
05/29/13 to 06/28/13 07/01/13 to 07/05/13 3480 3300 39270 -270 -270 6 580 5 0.6 0.6 830 1420 103.36
06/05/13 to 07/05/13 07/08/13 to 07/12/13 4070 3890 43160 -80 -80 6 678.3 10 0.2 0.8 830 1420 105.98
06/12/13 to 07/12/13 07/15/13 to 07/19/13 (600) (840) 42320 -1180 -1180 8 -75 10 0.2 0.8 830 1420 108.14
06/19/13 to 07/19/13 07/22/13 to 07/26/13 2880 2640 44960 -600 -600 8 360 5 0.4 0.6 830 1420 104.35
06/26/13 to 07/26/13 07/29/13 to 08/02/13 (110) (290) 44670 -1160 -1160 6 -18.3 10 0.2 0.6 830 1420 106.74
07/03/13 to 08/02/13 08/05/13 to 08/09/13 630 360 45030 -1300 -2420 9 70 10 0.2 0.8 830 1420 105.89
07/10/13 to 08/09/13 08/12/13 to 08/16/13 45030 5 3 2.8 830 1420 107.5
07/17/13 to 08/16/13 08/19/13 to 08/23/13 45030 5 3 2.8 830 1420 106.32
07/24/13 to 08/23/13 08/26/13 to 08/30/13 (190) (280) 44750 -320 -520 3 -63.3 5 3 2.8 830 1420 107.76
07/31/13 to 08/30/13 09/02/13 to 09/06/13 1970 1820 46570 -1060 -1060 5 394 5 0.6 1 830 1420 110.23
08/07/13 to 09/06/13 09/09/13 to 09/13/13 (850) (1000) 45570 -600 -940 5 -170 10 2 2.8 830 1420 108
08/14/13 to 09/13/13 09/16/13 to 09/20/13 1100 950 46520 -1800 -1800 5 220 10 1.4 0.4 830 1420 104.81
08/21/13 to 09/20/13 09/23/13 to 09/27/13 70 (50) 46470 -280 -550 4 17.5 5 0.8 0.6 830 1420 102.81
08/28/13 to 09/27/13 09/30/13 to 10/04/13 1620 1470 47940 -410 -410 5 324 5 1.4 0.6 830 1420 103.7
09/04/13 to 10/04/13 10/07/13 to 10/11/13 (1040) (1190) 46750 -1560 -1600 5 -208 5 1.2 1.6 830 1420 101.94
09/11/13 to 10/11/13 10/14/13 to 10/18/13 46750 10 1.6 1.6 830 1420 101.15
09/18/13 to 10/18/13 10/21/13 to 10/25/13 (1780) (1840) 44910 -1670 -1780 2 -890 10 1.2 1.6 830 1420 97.9
09/25/13 to 10/25/13 10/28/13 to 11/01/13 1010 920 45830 -490 -490 3 336.7 10 1.2 1.6 830 1420 94.6
10/02/13 to 11/01/13 11/04/13 to 11/08/13 (760) (910) 44920 -850 -990 5 -152 5 1.6 3 830 1420 95
10/09/13 to 11/08/13 11/11/13 to 11/15/13 (40) (220) 44700 -1400 -1400 6 -6.7 5 0.2 1 830 1420 94.39
10/16/13 to 11/15/13 11/18/13 to 11/22/13 980 800 45500 -550 -550 6 163.3 10 0.8 0.2 830 1420 94.72
10/23/13 to 11/22/13 11/25/13 to 11/29/13 45500 10 3 2.6 830 1420 92.78

Copyright © 2014 Dennis Meyers Trading Crude Light 5min Bars Using The Adaptive Goertzel DFT System Page 11 of 16
 In-Sample Dates Out-Of_Sample Dates osnp NOnp$30 NetEq ollt odd ont avosnp ncy pup pdn sTime xt CL CLS
10/30/13 to 11/29/13 12/02/13 to 12/06/13 2800 2650 48150 0 0 5 560 10 2.8 2.2 830 1420 97.82
11/06/13 to 12/06/13 12/09/13 to 12/13/13 (1210) (1360) 46790 -750 -1210 5 -242 5 2.8 2.6 830 1420 96.76
11/13/13 to 12/13/13 12/16/13 to 12/20/13 470 350 47140 -80 -80 4 117.5 5 1 0.8 830 1420 99.1
11/20/13 to 12/20/13 12/23/13 to 12/27/13 47140 5 1 0.8 830 1420 100.15
11/27/13 to 12/27/13 12/30/13 to 01/03/14 4030 3910 51050 0 0 4 1007.5 5 1.8 0.8 830 1420 94.31

Filter, LR<=3 and Top 5 mWb, Minimum mLTr

osnp = Weekly Out-of-sample gross profit in $
NOnp$30 = Weekly Out-Of-Sample Net Profit in $ = osnp-ont*30.
NetEq = running sum of the weekly out-of-sample net profits in $
ollt = The largest losing trade in the out-of-sample section in $.
odd = The close drawdown in the out-of-sample section in $.
ont = The number of trades in the out-of-sample week.
avosnp = The average out-of-sample profit per trade for that week $

Copyright © 2014 Dennis Meyers Trading Crude Light 5min Bars Using The Adaptive Goertzel DFT System Page 12 of 16
 Figure 1 Graph of Net Equity Applying the Walk Forward Filter Each Week
On CL 5min Bar Prices 1/06/2012 – 1/03/2014
Note: The blue line is the equity curve without commissions and the red dots on the blue line are new highs in equity.
The brown line is the equity curve with commissions and the green dots are the new highs in net equity. The green line is
the CL close price

Copyright © 2014 Dennis Meyers Trading Crude Light 5min Bars Using The Adaptive Goertzel DFT System Page 13 of 16
 Figure 2 CL GZ System 5 minute bar chart from 12/30/13-1/03/2014

Copyright © 2014 Dennis Meyers Trading Crude Light 5min Bars Using The Adaptive Goertzel DFT System Page 14 of 16
Figure 3 Partial output of the Walk Forward Metric Explorer (WFME)

The WFME Filter Output Columns are defined as follows:

Row 1 EC1FadmV is the strategy abbreviation, First OOS Week End Date(1/6/12), Last OOS Week End
Date(01/03/14), Number of weeks(#105) a=average of bootstrap random picks. s= standard deviation of
bootstrap random picks. f=number of different filters examined. c= slippage and round trip trade
cost(c=$30).
Filter = The filter that was run. Row 3 filter t5 mWb|lr3-mLtr

The t5 mWb|lr3-mLtr filter produced the following average 105 week statistics on row 3.

tOnp = Total out-of-sample(oos) net profit for these 105 weeks.

aOsp = Average oos net profit for the 105 weeks

aOTrd = Average oos profit per trade

aO#T = Average number of oos trades per week

B0 = The 105 week trend of the out-of-sample weekly profits

%P = The percentage of oos weeks that were profitable

t = The student t statistic for the 105 weekly oos profits. The higher the t statistic the higher the probability
that this result was not due to pure chance

std = The standard deviation of the 105 weekly oos profits

llp = The largest losing oos period(week)

eqDD = The oos equity drawdown

lr = The largest number of losing oos weeks in a row

# = The number of weeks this filter produced a weekly result. Note for some weeks there can be no
strategy inputs that satisfy a given filter's criteria.

eqTrn = The straight line trend of the oos gross profit equity curve in $/week.
rd
eqV^3 = The ending velocity of 3 order polynomial that is fit to the equity curve
2
EqR2 = The correlation coefficient(r ) of a straight line fit to the equity curve

Dev^2 = A measure of equity curve smoothness. The square root of the average [(equity curve minus a
2
straight line) ]

Blw = The maximum number of weeks the oos equity curve failed to make a new high.

Copyright © 2014 Dennis Meyers Trading Crude Light 5min Bars Using The Adaptive Goertzel DFT System Page 15 of 16
BE = Break even weeks. Assuming the average and standard deviation are from a normal distribution, this
is the number of weeks you would have to trade to have a 98% probability that your oos equity is above
zero.

eff = Efficiency. The average daily out-of-sample profit divided by the average daily in-sample profit.

tOnpNet = Total out-of-sample net profit(tOnpNet) minus the total trade cost.
tOnpNet=tOnp – (Number of trade weeks)*aOnT*Cost.

Prob = the probability that the filter's tOnpNet was due to pure chance.

Copyright © 2014 Dennis Meyers Trading Crude Light 5min Bars Using The Adaptive Goertzel DFT System Page 16 of 16