Module 4, Lesson 2
TradeStation Walkthrough
Tomas: Okay, so let's start with the visualization indicator again.
So, you have a simple indicator here in this Module 4. And what it basically does, or
how we use it here for demonstration purposes, so in this window I took the VXB.X
Volatility Index, and inserted it as a second chart, and I applied simply this indicator,
which now splits into three different zones.
So again, we have Volatility Zone 1, 2, and 3.
And then here, below, we use the dependent input, that means Average True
Range, instead of independent input.
So it's basically still the same thing. We are using two different inputs. One is inde-
pendent Volatility Index; the second one is dependent, Average True Range volatility
computation or indicator; and we again split them into three different zones, be-
cause behaviour of our trading strategies and entries will definitely be quite differ-
ent within each volatility zone.
So here is the indicator.
We have the first version, is the independent volatility input. That's the VXB.X, and
the second one is Average True Range.
Again, I highly recommend staying with one period, which is 25 bars. Do not change
it. Keep it the same. We don't want to do any overfit, over optimization. Just use this
as a fixed period among all your strategies you're ever going to test with the dy-
namic position sizing technique.
And, again, we just take 100 bars, or 100 days, and split into three zones.
So here is the optimization code.
Now, notice that in this code there are a couple of important new things.
So, first of all, we will also use DATA 4 and 5.
Page 1 of 6
(c) 2018 [Link] / Distribution without author´s approval is not permitted.
TRADING INVOLVES A SUBSTANTIAL RISK OF LOSS AND IS NOT SUITABLE FOR ALL INVESTORS,
PAST PERFORMANCE IS NOT NECESSARILY INDICATIVE OF FUTURE RESULTS
4 will be VXB.X, same ... oh, that's my mouse. Sorry. Jumping a little bit.
So DATA 4 will be VXB.X. The same time frame as you're using your underlying strat-
egy in DATA 1.
And number 5 - data feed 5, will be VXB.X daily.
Okay. All data must be set to local in TradeStation. I'll show you.
And then it's basically what you know already.
So the first technique is based on the independent volatility input, based on 100
bars.
Second one is based on Average True Range, calculated on the main time frame.
Then we will use higher time frame again. Daily time frame for calculation of the in-
dependent volatility input – the VXB.X – which we will use 100 days again.
And, second one, we will use daily chart, which we use Average True Range, period
25, and, again, we will split it into three zones.
So you all know this from the previous modules already. So you know already how
and where to insert your strategy code, strategy exits, and how to modify the strat-
egy code in the entry part.
So let's get to TradeStation now.
So here we can see that we have five different data inputs. That's the underlying
strategy. If your strategy uses DATA 2, it will be here. And then we need daily data
from your underlying market to calculate Average True Range on daily data.
And then we have VXB.X, and we have the same time frame, like underlying market
and daily data, so we can try all four techniques.
And now do not forget that we need to set local time zone everywhere, so if you are
using time in your strategy - time filter, or T-segment concept - you need to change
the times of your strategy into local time. Everything needs to be in your local time,
not in the exchange time, otherwise you will not be able to do this properly, so it's
very important. Everything needs to be set into your local time, which definitely will
be, in most cases for most of you, different than my Spanish time.
And that's the only thing that you need to be aware of.
Page 2 of 6
(c) 2018 [Link] / Distribution without author´s approval is not permitted.
TRADING INVOLVES A SUBSTANTIAL RISK OF LOSS AND IS NOT SUITABLE FOR ALL INVESTORS,
PAST PERFORMANCE IS NOT NECESSARILY INDICATIVE OF FUTURE RESULTS
Okay, then we do the usual optimization which you already know from previous
modules - which is again, optimizing technique from 1 to 4.
And here we're optimizing between 1 and 3 contracts per each zone, differently for
long and short.
We set 20% of data as out-of-sample, and, of course, we must not forget 100 bars
back to reference to set it up here.
And we run the optimization.
The optimization would last about an hour, so I've done it in advance already, so….
Oh, I've just lost the Optimization Report because it was recalculating right now, but
that's okay. I have a screenshot, so let me show you.
Okay, so here is the screenshot of the Optimization Report as you get it when you
use the set up as I just discussed.
So, again, you chose in-sample, sorted by your fitness function - in this case it would
be TradeStation Index.
Now you need to use the same fitness function that you used for your strategy. So
the same fitness function that you use when developing your strategy, which you
are going to use with dynamic position sizing. You need to be consistent with this
and keep using the same fitness function all the time.
And when we have a look at the first 20 best results, we can see that most common
is solution 3 - Technique 3 and Technique 4.
Technique 4 is more common than Technique 3. I think there are way more occur-
rences of Technique 4 than Technique 3.
So we would use Technique 4, and first available on the line number 2 shows aver-
age trade 81.46, which is very good, among the highest average trades we can see.
So that's the one we would use.
And we will have a look at in-sample and out-of-sample again and see the results.
Let me get back to TradeStation.
I'm not really sure if I will be able to see the in-sample, out-of-sample results now
separately, because….
Oh yes, it’s all here. Okay, so let's have a look.
Page 3 of 6
(c) 2018 [Link] / Distribution without author´s approval is not permitted.
TRADING INVOLVES A SUBSTANTIAL RISK OF LOSS AND IS NOT SUITABLE FOR ALL INVESTORS,
PAST PERFORMANCE IS NOT NECESSARILY INDICATIVE OF FUTURE RESULTS
So it's Technique 4 and we have in-sample here, so let's have a look if we have rea-
sonable out-of-sample.
That's quite alright. Okay, yeah, we had this small drawdown, which got it a little bit
to minus; but from here, a little bit of stagnation and then a growth, so it looks al-
right.
It definitely, if we plot the entire equity curve, it's just continuous more or less
within the tendencies. So that's quite alright. I wouldn't mind this one.
Okay, so Technique 4 – be it. That's the final one, and the number of contracts will
be 2, 2, 3, 1, 2, 1.
So that's what we're going to use in the result code: 2, 2, 3, 1, 2, 1.
Technique 4.
So the winner is Average True Range on daily data - which is quite independent. It's
partially independent because if you're underlying strategy does not use daily data,
then even if we're using this dependent Average True Range input, we're using daily
data so there's a little bit of independency, which is definitely better than if it would
be the same underlying time frame. That's quite alright.
And also you saw that the second biggest winner was case number 3. So you can try
this one as well if you want to.
So the second winner would be Technique 3 and contracts would be 2, 3, 3, 1, 1, 1,
2.
2, 3, 3, 1, 1, 1, 2.
3, 3, 1, 1, 1, 2.
I think it was Technique 3 and, of course, you can have a look what it looks like. And
it's alright if you use this one as well.
So the first one or the second one in the optimization time frame, that would be
quite all right. I probably would go directly for the number 2 without any further
look, but this could be pretty good as well.
So that would be the in-sample.
Out-of-sample?
Yeah, it all looks good. Looks quite alright. No problems here.
Page 4 of 6
(c) 2018 [Link] / Distribution without author´s approval is not permitted.
TRADING INVOLVES A SUBSTANTIAL RISK OF LOSS AND IS NOT SUITABLE FOR ALL INVESTORS,
PAST PERFORMANCE IS NOT NECESSARILY INDICATIVE OF FUTURE RESULTS
The equity is, in general, smoother, so you could go with the fully independent input
with this version as well, and you definitely would not make a mistake with this one
neither, okay?
So you can definitely use this one as well.
However, I would choose the Technique 4 because there were more occurrences in
the optimization profile and here we would have the final result.
So you have all the codes there and let's compare the final results.
Okay, so here we go.
So net profit would jump from $25,000 to almost $60,000. That's a 138% improve-
ment.
Now, maximum drawdown, it has increased quite a lot - by 47% - which is higher
than I would like to see, so maybe in this case I still would have a look at Technique
3.
So what would be the Technique 3 ratio here?
Technique 3: $66,000 and $7,500 drawdown - so that would be definitely higher
profit. 66 versus 60. And definitely lower drawdown. Here we have $8,600, before
that we had $7,600.
So that would probably be a better solution. I would probably go with 3.
Also, it's independent, which is definitely way better. The equity was smoother.
But if we stay with this second one, which is little bit worse, we still are improving
the net profit/drawdown ratio by 63%, which is very, very, very nice.
But let me do a quick calculation with the third technique.
It was about $66,000 profit, and it was $7,600 drawdown, so that would ratio 8.7.
So with the second technique, actually, with Technique 3, which would be the inde-
pendent VIX on daily data input, we would improve the ratio - let me do the quick
calculation - by about 100%.
Okay, so from 4.3, we would jump to about 8.7 so by 100%, which is much better.
That's what I'd like to see, so that would be a better one definitely.
Page 5 of 6
(c) 2018 [Link] / Distribution without author´s approval is not permitted.
TRADING INVOLVES A SUBSTANTIAL RISK OF LOSS AND IS NOT SUITABLE FOR ALL INVESTORS,
PAST PERFORMANCE IS NOT NECESSARILY INDICATIVE OF FUTURE RESULTS
But this one is not bad as well. You can see that with both of them we have - some-
times we have more options, and with more options we still can, with both of them,
get a pretty nice result and improvement.
Okay, so that was the volatility-based dynamic position technique. I hope you like
this one and in the next module we will cover market internals techniques for an-
other dynamic position sizing approach.
So, see you in the Module 5.
Page 6 of 6
(c) 2018 [Link] / Distribution without author´s approval is not permitted.
TRADING INVOLVES A SUBSTANTIAL RISK OF LOSS AND IS NOT SUITABLE FOR ALL INVESTORS,
PAST PERFORMANCE IS NOT NECESSARILY INDICATIVE OF FUTURE RESULTS