Up next in 10
Marsten Parker - Lessons from Systematic Market Wizard - AlphaCast Ep. 1 - 2023-02-23
Dec 11, 2024
Welcome to the first episode of AlphaCast! In this episode, we are thrilled to have a special guest, Marsten Parker, the only market wizard in the world of systematic trading.
Marsten shares his journey of getting into trading and the steps he took to become the trading legend he is today. He reveals the trials and tribulations he faced in his early years of trading and how he developed his own unique strategies that led to his success.
With Leo's probing questions, Marsten takes us on a deep dive into his trading strategies, both from his early days and those he still employs today. He discusses the importance of strategy diversification, why it is crucial to his success, and how he adapts to market changes.
Throughout the episode, Marsten shares his wealth of knowledge and trading insights with us. Whether you are new to trading or a seasoned investor, you will find something valuable in this episode. So sit back, relax, and join us as we explore the fascinating world of trading with the one and only Marsten Parker on AlphaCast!
✉️ Learn the secrets of legendary traders and investors:
https://bit.ly/4aCna7i
🔗 Follow along:
https://analyzingalpha.com/episode-001-marsten-parker
⏱️ Video Chapters
0:00 Introduction
0:50 Marsten’s early days
4:40 Marsten’s first strategy
8:30 Origin of RealTest
9:30 Becoming systematic
14:40 Initial breakout strategy
17:30 Eight is great
20:45 Finding new ideas and Stockbee
25:00 Position sizing
26:30 Regime filtering
28:30 Key lessons
30:20 IPO strategy
36:15 The devil is in the details
38:30 Systematic vs discretionary
43:02 New idea to trading strategy
47:20 Risk management using margin requirements
48:28 Marsten’s advice for new and experienced traders
52:00 RealTest explanation
#trading
#systematictrading
#algorithmictrading
#technicalanalysis
Website: https://analyzingalpha.com
Interested in backtesting and code? Check out my sister channel:
https://www.youtube.com/channel/UC3iZRb-i8GwMv3FE9cY95hw
** All videos and code examples are subject to the disclaimer **
Disclaimer: https://analyzingalpha.com/disclaimer
Show More Show Less View Video Transcript
0:00
welcome to the very first episode of alpha cast a show dedicated to interviewing Traders and investors
0:07
who've managed to create Alpha in this very first episode I have the absolute pleasure of interviewing Marston Parker
0:14
for those of you that don't know Marston Parker is the world's only systematic
0:19
Market wizard whose trounce 99 of New York hedge fund managers for over two
0:24
decades in this interview we start by understanding how he got into investing his journey to become a systematic
0:30
Trader and what he's up to today and during the interview he shares some insights that you're probably going to
0:36
be hard-pressed to find elsewhere so without further Ado Marston Parker Marson a pleasure to have you speak with
0:44
me today the first thing that I'm really interested in and just kind of understanding your story you know can
0:49
you give us a little bit of a background about how you you know got into trading and and you know that path
0:55
sure it's a pleasure to be here thanks for inviting me on your uh your new
1:00
podcast and um of course so my um my story uh in a nutshell
1:08
um going back to high school my two main interests were classical violin classical music and computer programming
1:15
um I actually pursued the former by going to a conservatory in New York City uh it
1:21
was an efficient way to learn that I wasn't going to be one of the top violinists and and that was in the early
1:27
80s when um personal computers were first coming on the scene and that was an exciting time and I just got more and
1:33
more into that um moved back to Boston where I grew up and uh managed to get myself a computer
1:39
and um learns program and very luckily without having any formal education in it got
1:46
offered a job as a software engineer and you know moved among companies the
1:52
other long story short um the the last company I worked at
1:58
actually went public in 1996 and um I owned about four percent of it
2:05
um again in those days companies went public and they were tiny so four percent you know it was only like a
2:11
million or two but um you know so I had a good good exit and then I I had an
2:16
investment problem of what doing that money and I didn't want to just give it to a financial planner and pay their fee
2:23
and everything and so I started buying books about trading and investing I started out thinking investing but
2:29
kind of gravitated more towards shorter term trading um sure yeah I just dabbled in this and that
2:36
discretionary stuff for about a year and lost some money and then um uh met somebody who was who was writing
2:45
about a systematic approach on a website called thestreet.com back then and ended
2:51
up getting to know him and working with him for a while and
2:56
he went from there um okay now did you ever give your money to a financial advisor by any chance
3:03
no I I have a meeting with one and yeah so now I don't I don't think I'll take
3:09
this path because you're you're smarter than I am and I do it myself yeah you
3:14
know what's funny is because I ended up um you know I I kind of came from a trading background from a completely
3:20
different perspective I started with valuation and and whatnot but that's the how I got into uh investing initially
3:27
and then trading uh was um you know I I purchased some shares
3:32
and then I you know the stock went up and then I looked at my account balances and I still somehow lost money and then
3:37
that was a clear message when I looked at the fees I'm like you know what I'm gonna have to start figuring out how to do this stuff myself yeah
3:44
you know obviously I know a lot about your background I've read the market Wizards uh books and things like that
3:51
and I guess you know the the first question that comes to mind um is you know back in those early days
3:57
you know whenever you were constructing uh these trading systems I guess you first started learning technical
4:03
analysis with I guess it was your your one partner initially and then you started back testing uh can you give us
4:08
some ideas of maybe some of those strategies that you were using back then to kind of give the audience who uh a
4:15
lot of the Watchers on this channel are interested in trading and basic systems to kind of give them an understanding of maybe what what those systems look like
4:22
sure I mean when I first started I didn't I didn't have any techniques and I
4:27
didn't even know what techniques I was supposed to be looking for um as I recalled it but but I kind of
4:32
knew from the beginning just because of my personality I wanted something systematic in other words I wanted a set of rules to follow I didn't want to just
4:40
try to do it purely by intuition um and um as I
4:46
I think the very the very first um rule I came up with in a book
4:51
uh was um p uh to you know take take
4:56
like the stocks in a major index and rank them by pe overgrowth and uh choose
5:02
the ones with the highest PE divided by growth rate um or no the lowest p e divided by
5:09
growth rate yeah it was sort of like uh you know growth at the right price I
5:14
think that was even used as an acronym back then um you know and I remember putting on a
5:20
couple of positions like that and holding them for a week or two and seeing they didn't go anywhere and then I got bored and
5:26
decided to try something else so I had I had no real conviction but I was reading
5:32
wherever I can't code online and in books and like this guy Gary B Smith was writing a column called technicians take
5:38
in the street.com and roughly what year was this in 1997.
5:43
okay he um yeah I think you can still find his
5:49
articles on there uh one or two of them even mentioned me but um he um
5:55
he was unusual among things you could find online and that he he just disclosed the exact rules of the
6:02
strategy he was trading and it was kind of it was rule-based with some discretion on the entry side and
6:08
completely mechanical on the exit side and it was a derivation of the William
6:13
O'Neill um strategy and how to make money in stocks there's so-called canceling but
6:19
subtracting out all the fundamental parts and using a smaller profit Target
6:25
um but so base you know the basic the daily process of that system was run
6:30
some kind of a scan for stocks that had unusual volume breakouts the prior day I
6:36
mean initially our scan was was you know to literally look in the printed edition of Investors Business Daily because they
6:42
had a table of them and then type type those symbols into a charting program and see if it looked like nice tight
6:49
consolidation prior to a breakout uh yeah and then simply
6:55
by a fixed number of shares at Market at the open the next day and tell the
7:00
broker to put a bracket of a I think we used a five percent Target and a seven
7:06
percent stop or something like that um and then you just don't do anything so one of those is hit
7:12
um and and so I started trading that strategy with him in early 1998. you know and there was also a short version
7:19
as well so we were always trading alongside on the short side um and then in hindsight that's probably
7:25
the most important thing I got out of that was to was to always trade both sides um it was my first foray into strategy
7:33
diversification um and and we had a we had a good year it
7:40
was something like plus 50 in 1998 and 98 was not like 99 it was a it had big
7:46
moves in both directions um but you know we were able to catch
7:51
pieces of the moves in both directions because we're long and short and um but and so that was the semi
7:57
mechanical semi-discretionary approach that year and then I got and then during that year I was starting to build the
8:04
back testing just because I was you know I would look at the trades we were doing and I would be curious but what if we
8:09
had used these exit rules instead of those and and then so first I built something to
8:14
test the exit rules and then I started wanting to test the entry rules and then I started wanting to test different
8:19
entry rules so it just kind of evolved uh bottom up like that from what we were
8:25
actually doing and is that I know that you have real tests is that the initials uh that
8:32
origin of real test yeah I think okay I think my first version of it was scan
8:37
test you know because our process was eventually we we realized we could get data over the telephone through a modem
8:45
and run scans instead of having to look in the newspaper um and then I started wanting to test those scans for the
8:51
first version okay for those of you that don't know real test uh is Marston's
8:56
back testing engine that you know from a unique perspective it can test multiple strategies at once which you know help
9:02
significantly with the equity curve right okay and we'll we'll talk more
9:08
about that and I'll put links in you know the description below about your software um but uh you know so then I guess from
9:14
you know after after that I think that you you choose went sort of separate ways because his he was really focused
9:21
on the technical analysis and I think you like me you know whether For Better or Worse right we can talk about trading
9:28
intuition later uh like the the concreteness of of systematic trading right I mean uh through his position at
9:36
the street.com Gary was able to start a subscription service called chart man I think it was
9:43
um and um you know basically uh sell trading signals that way and um you
9:50
know and he he took pride in his ability to look at a chart and you know like look
9:56
at several charts and say this one's better than that one and and I never got any sense that I
10:02
if it was anything there for me to me the distinction between two similar looking carts was completely arbitrary
10:08
and when I tried to print you know parameterize the differences in back system like that validated it and that
10:16
is at least at least for me I mean maybe there's subtle nuances that somehow at Edge but I seriously doubt it
10:22
um and um so so yeah I mean we we stayed in touch
10:28
and he I don't know if he still is today but he kept using real tests for years uh for for some of his own stuff
10:36
um and um and I I I start I've I started uh first
10:44
100 mechanical trading at the start of two thousand I was kind of um moving towards it in 99
10:51
but I I kept doing discretionary interventions and discovering that they made things worse and um
10:59
sure so essentially you added you know algorithmic you know order execution is that essentially what you mean
11:06
no I meant I mean that I I removed all all the discretion and and deciding what
11:12
stocks to buy or where or where to exit you know that didn't allow me myself to intervene in that I did have
11:20
order automation starting in 2000 but I mean in the sense that but we were using a small uh
11:28
so-called agency broker in New Jersey at the time and um you know when we first
11:33
started we would actually call them every morning with our orders on the phone and then um
11:38
well we figured out we could email it to them so so so my automation was actually
11:44
to calculate what what I want to buy and what or sell and what sizes and then
11:50
automatically generate an email and automatically send it to them and that that happened before the open
11:58
every Market Day later
12:03
um over you know after a couple of years of that they um especially in 2005 when I started
12:10
generating orders throughout the day not just before the open uh they gave me a
12:15
an interface like they had a fix interface but I didn't have software deal with that I didn't want to learn
12:21
the whole fixed protocols but they gave me a little Windows app that that was that could connect to their fixed
12:27
interface and I could drive that Windows app from my my
12:33
automation which was implemented in Excel at the time Excel on VBA
12:38
um so so that'll work and then I think 2013 maybe I changed to interactive brokers and and redid everything using
12:46
their API so you had uh 23 was that 2013 or well you had 14 years of consistent
12:52
profitability then you had three years where you almost moved in the town or through in the towel and then you had
12:58
four years that roughly um you know some of the the best years ever so you know kind of take me through
13:04
that Journey um of you know how you uh you know that 14 years consistent profitability is
13:10
that you were using the the system that you're talking about now I was pretty yeah those the whole time
13:17
from 2000 through um 2013. I was just running two
13:24
strategies one long and one short I mean I had many different versions of those two strategies in terms of changing
13:30
parameters and changing the rules slightly but but it was just basically the same thing that I had implemented
13:36
with with Gary back in 98. uh you know and then mechanized in my own way um
13:43
uh and and you know those they work great until they didn't I mean it wasn't
13:49
just throughout that time between 2000 and 2013 there were a couple of different times when they seemed to stop
13:56
working and I made changes um uh particularly in in early 2005
14:03
you know basically when when I hit a hit a drawdown that was worse than I had had
14:09
before or expected I would retest and try try to find new ideas and it's not
14:16
necessarily the best practice but I think it's pretty much what everyone does you know you have a bad drawdown and you tweak your system to get rid of
14:22
that drawdown and and then you resume uh and and hope that that was somehow meaningful
14:29
um so you know it's one of those things you hold your nose and do even though you know it's probably bogus so so yeah
14:35
I did that a few times in that that period um yeah and the the biggest change I made
14:40
was in you know since I was my the main point of my strategy was to
14:46
look for a breakout from consolidation with unusual volume uh now this was before this is before
14:52
you were in stock beat correct correct I joined doc B when when my trading was started
15:02
to go back I mean in 2013 uh when it seemed like it wasn't working anymore
15:08
especially the short side um when I first started in 2000 uh you
15:13
know clearly that was a good year to have a have a short strategy and and I made all my profit that Year from the
15:18
short side and it was like yeah all the although some of the people here might not be old enough to remember what
15:23
happened there that's the uh it's true the.com.com album and but I should say
15:29
right but yeah and that's kind of got implanted in my psyches okay I need to I
15:36
need to always trade the short side and and you know it's when you had good initial success with anything it's hard
15:42
to let it go and uh you know 2012 2013 is when the the um
15:51
I don't know everybody realized just buy the dip just buy the dip you know that funny YouTube video yeah yeah btfd if we
15:58
win it yeah right yeah um and and so you know that the the shorts
16:04
the strategy of what I had found previously worked well is the bigger the breakdown you know you have a stock that's been
16:11
going up and then suddenly it has a huge breakdown usually it was like a negative earnings morning or something like that
16:17
and you could just short it after that big break on it would keep dropping for days that we don't practically always work
16:23
but but in late 2012 the regime shifted where that became a Buy Signal uh you
16:30
know and it it took me a whole year to admit that that change had happened and and a pretty big drawdown
16:36
um and that's when I added I finally said okay I'll do the opposite and I got into mean
16:43
reversion trading but um I do I mix things up now I run some some
16:51
multi-day meaner versions and one day meaner version some short-term Breakout
16:57
uh and breakdown in fact I've just I've just now reintroduced a new version of
17:03
my old strategies which I hadn't traded since 2015. um because with the rising interest
17:09
rates it seems like we might be back in more like that older regime again sure
17:14
sure so I get uh so roughly how many uh strategies are you trading right now I'm running eight
17:22
um so yeah there's a four Mainer version and four I'm going to say Trend following but
17:28
it's not Trend following like a Futures Trader would use it it's it's like a it's still a one to five day swing I I
17:35
like that you know one to five day time frame I just haven't found things that have
17:42
the return versus drawdown ratio in longer time frames
17:48
right and then um I guess briefly you know why would you want to trade Aid
17:54
strategies consecutively um be well because they they do
18:00
different things at different times I mean you look for for them being not correlated with each other I mean it's
18:05
just an extension of what I was doing from the beginning by trading two instead of one and specifically long and
18:11
short I want it to be Market neutral and and then and then I thought of it in in
18:17
you know adding another dimension besides side is sort of um you know
18:22
style or factor um you know professional
18:28
hedge fund managers understand this much better than I do but you know you're always looking for different factors and
18:33
multi-factor models and that sort of thing so I do my own little amateur version of
18:39
that have you ever seen the ray dalio Holy Grail video where he shows the the
18:46
returns of uh you know different strategies that are uncorrelated by any chance
18:51
I don't think I've seen that one I'll have to look for I mean I've certainly seen a bunch of his stuff and read his books but I I'm not familiar with that
18:58
video off the top of my hand okay so I'll um so basically in essence it just shows how powerful uh combining multiple
19:05
strategies can be right well I should say multiple uncorrelated strategies um and that's essentially what what
19:11
you're doing and again what your your real test is is really good yeah yeah awesome yeah and they just have to be a
19:17
little bit uncorrelated they you're not going to find perfectly negatively correlated or
19:24
yeah yeah it's funny because I learned that in the uh the covid uh the covert
19:30
crash right uh one of the things that that um you know obviously when you know stocks go down you expect uh bonds you
19:36
know historically go up but that's not necessarily the case whenever uh you know the liquidity dries up everything
19:42
just Falls right so you have to really think careful about um you know is this really you know
19:48
negatively correlated or is it depending upon the market regime I mean I guess the only way to really be negatively
19:53
correlated would have been it is short but but you get the idea but I first discovered that phenomenon you're
19:59
talking about in I think it was August of 2007 that there was an event that's
20:05
known as the crunch I think uh and um you know I noticed all of a sudden both
20:12
my lungs and my shorts were losing money at the same time and that was very
20:18
confusing and I didn't understand it until like a year later when people wrote papers about what the mechanics were basically it was
20:25
a liquidity event yeah yeah yeah that's uh yeah definitely something that uh that that I learned
20:31
too I mean I still did all right but um could have done a lot better uh you know next question would be you know how is
20:36
your trading since you know that 2013 you joined Stock B started doing mean reversion
20:42
um you know can you tell us a little bit about you know how you uh your Evolution from there and and the different trading
20:48
systems uh maybe give us one you know historically you know nothing that you're using active uh that maybe the
20:55
audience could uh you know get a better understanding of what we're doing in today yeah well so I when when my
21:03
results started to go south in 2013 I was kind of looking far and wide for new ideas and so on and I somehow stumbled
21:10
on stalked me on an internet search and noticed that they seem to be um
21:16
also trading you know swing systems based on O'Neill's work so I thought oh
21:22
here's somebody who's doing what I do and maybe it's still working for them so that I'd go in the community on that
21:28
basis um and um you know but it's it's pretty
21:34
much uh discretionary approach to that and and over time you realize it's mostly about knowing what stocks are in
21:41
play on any given day and uh having good situational awareness and all the stuff that good discretionary Traders do
21:48
um and so strangely even though nobody on Stock B was doing meaner version uh
21:54
somebody mentioned it to me on there um and which I I wasn't even familiar
22:00
with the concept really or and so I started looking into it uh you know like
22:06
found the books by Larry Connors that describe his his approach um
22:11
and you know the the basic idea with the meaner version strategies is is um
22:17
you're looking for like if you want to buy a stock that's been going down for a few days and then you buy it by placing
22:24
a limit order if it goes down even more tomorrow so so you're looking to basically catch
22:31
a falling knife at a reversal Point um and you have to uh you know keep your
22:37
sizes smaller when doing that uh but you know when I when I first became aware of the the strategy concept and started
22:44
testing it I you know the results are so much better in back testing uh than
22:50
anything you can find with breakouts so it really surprised me um
22:55
so it but I mean I've always been pretty cautious and conservative in terms of adding things adding new approaches or
23:02
changing what I do um and um so that the first year out
23:08
2014 I I used half my account for my
23:13
existing system and half for um meaner version and the the
23:19
the the meaner version and I I always insisted having a long short pair so from the beginning I also had a short
23:26
year version strategy um yeah the meaner version pair
23:32
you know in its own terms had a plus 40 year so in my account terms was plus 20
23:37
and the other the um what I call my classic strategies made
23:43
zero um and then so in 2015 I said okay I'm just going to trade the two meaner version strategies they obviously are
23:49
brilliant um and um so you know basically doubled my position sizes and those
23:55
and um and the year started amazingly uh but
24:01
then uh there was some very serious volatility in the market in the second in the summer of 2015 and continuing
24:08
basically from around mid-2015 into the start of 2016.
24:14
um yeah the long side in particular just got destroyed and I
24:20
did and I kept in I had already discovered with with the meaner burden strategies they have a nice
24:26
characteristic when they have a drawdown they tend to bounce back really quickly and and so I
24:32
I had backtested them with you know not reducing your size when you get into a
24:37
drawdown because then it bounces back faster so that seemed like a good idea so I was trying that in 2015 late 2015
24:44
during the the extreme volatility and uh kind of got my head handed to me as a result
24:51
um so I learned learned the hard way about keeping your size really small with um
24:58
with a meaner version and the other thing because with my breakout systems I had always used the position size method
25:05
that Gary had preferred which is to trade literally the same number of shares on every position whether the
25:13
price was 10 or 80 or 50 or whatever um you know I mean we didn't
25:18
trade go up to I think back then there weren't very many stocks that were over a hundred it would seemed much more the
25:25
case before 1999 almost all the stocks lived in that kind of 20 to 80 range
25:30
and it actually is a reasonable method because the ones towards the lower end of that range have more volatility on a
25:37
percentage basis so you end up sort of volatility adjusting yourself if you just trade
25:42
fixed shares because you trade you know smaller dollar positions at the lower end of the price range and back then
25:49
that worked now I would not recommend it but I applied that same I figured let's
25:54
develop newer version strategies with that same sizing technique but that was not a good idea
26:00
um because because you know I ended up having some positions that were like maybe up to 20 of my account in one
26:07
position and then getting it trying to catch a falling knife at that size
26:13
you know uh so that wasn't so good so yeah that I adjusted to that okay now
26:21
I know you said that you had um you know some of the strategies uh in essentially had challenges during various periods do
26:27
you do any like regime filtering or anything like that or your strategy is always on
26:34
uh they're still always on I mean I've experimented with with dream filters I do use
26:40
you know long-term Trend filters for the individual stock and some of them not in all of them but um
26:48
uh every time I try to back test a general purpose regime filter I find it
26:53
hurts the results now that's probably because if you take a system that you've already optimized and then apply any
26:58
kind of filter to it it'll probably make it worse you almost have to start with the idea of using a filter if you want
27:05
to use a filter but um anyway that's that's what I do no that makes sense
27:11
yeah so for instance your uh you know you mentioned a trend there so uh you know your your long mean reversion I'd
27:17
assume there's probably some type of Trend in there you want to buy stocks that have been going up and then had a short down period would that be a
27:24
correct assumption yeah right yeah you usually um
27:30
it's good to just have a long-term uh you know like above the 200-day moving
27:35
average let's say um and that that that that gets you to not trade the strategy or
27:41
I mean during a during an extreme Market meltdown like 2008 hardly any
27:47
stocks are above their 200-day moving average so you just end up not getting any entry signals so it kind of acts as
27:52
a regime filter on its own but I mean the thing with with a shorter holding period regime filters let at the
28:00
market level is less relevant because there's always individual stocks moving in both directions on a shorter time
28:05
frame um so you just end up leaving too much on the table with a routine filter at
28:11
the market level you know for a longer term strategy like a monthly rotational NASDAQ momentum system or something you
28:18
definitely wanna want it in there okay no no that's interesting okay
28:24
perfect and then I guess um you know what is uh you know from there what does your your trading look like today maybe
28:29
some of the you know what have you learned like maybe some of your key points that have you learned you know over your over your time period Trading
28:38
um the the main directions I've been I've evolved besides what I was just talking about in terms of how to size
28:44
positions um and now I always just use a simple fractional sizing right now I'm
28:49
I'm using five percent of my account as my position size on air on every trade
28:55
and it as I add more strategies I'll reduce it I was using ten percent when I
29:00
only had four strategies and uh and now I'm using five partly because sometimes two strategies trade the same symbol so
29:07
I want to keep the same stock exposure lower um so
29:13
more conservative sizing diversification with more strategies and um
29:19
simplification of the strategies I used to add too many rules uh
29:25
try too many parameter values and and I I just keep um
29:31
keep simplifying them more as I go along I'm just trying to I'm
29:38
more focused on just wanting to you know there's a certain Factor you want to capture it's a short-term meaner version
29:44
short-term breakout whatever um foreign I've also started the the what's allowed
29:52
me to because I had there's basically a two by two Matrix right you have long versus short for one
29:58
axis then you have momentum versus meaner version on the other and so I kind of had that covered with my four strategies for a few years there like
30:06
2016 through 2018 or so and then I um thought okay well how can I add to this
30:13
and and I got different subsets of the universe so I actually this was another
30:18
idea I got from from not only from stockby but from someone who you featured in some of your
30:24
videos who goes on Twitter by Scotland um he was very focused on trading IPOs
30:31
in Stock B uh this was I think before he got really active on Twitter he was in
30:36
stock sharing these ideas so I I came up with um
30:42
with a system for that for trading recent IPOs on a you know when they break out to a new all-time high and um
30:49
and that did really well for me from like 2017 through
30:55
through the middle of 2021 hasn't done anything since then because there haven't been very many but um
31:03
you know and that one kind of has its own built-in regime filter because there aren't any IPOs when the Market's going
31:10
down right no that's interesting yeah that's true you know it's a race race Capital unless the Market's frothy right
31:17
they want to get there as much as they can you know and then you can also I mean I use the nor gate data which which
31:23
has all the historical index constituency uh available and so you can
31:29
experiment with different sub universes that way you know anything from
31:34
stocks in a specific index like Russell 2000 only or to um stocks that aren't in any index
31:41
um it's another interesting one to look at mostly you get adrs when you do that but
31:47
okay great yeah um that's the data provider that you use
31:53
for your back testing nor gate yeah yeah no if they're they're pretty good I'd actually did an analysis of
31:59
some of the various providers and Norgate uh you know was one of the was one of the better ones yeah they're very
32:06
good I mean it's all they do and they because what you really want when you buy data is is also metadata you know
32:13
you you I mean especially if if your trading involves uh starts with saying I
32:19
want to scan the whole stock market every day for and you know with like a liquidity filter to find the best uh
32:25
setups um you need someone to maintain that definition of the whole stock market you
32:31
know what is that what are what are all all the symbols for U.S Common Stocks
32:37
and so on and be able to know something about each one uh you know at least like
32:42
it's um what exchange it trades on what index it's in on you know that sort of thing
32:48
and it's way too much work to maintain all that yourself from free online sources it's worth paying
32:55
for data for sure yeah then the survivorship bias and you know let me see your your lungs all look look better
33:01
right because you know nothing uh you know is in the index anymore that that fell out so yeah yeah
33:07
besides the survivorship bias which I mean one is to especially for longer term strategies
33:13
you want to make sure your data includes things that died like Enron and will come but also
33:20
uh conversely like if you're if you start with the universe of everything that's ever been in the Russell 3000 and
33:27
you just trade that uh you get look ahead bias because you're trading things
33:32
in the past that are going to be added to the Russell 3000 later um right because they grew enough to do
33:38
so so um so that's why the historical constituency is so important right right
33:43
no absolutely perfect okay um so then uh you know today I mean
33:48
how's how's trading going today uh for you Larson give me this literal day or you mean
33:54
recently oh just in general yeah yeah yeah this very day right now I think I'm
34:01
actually having an update today because uh okay I generally do better when the Market's going down uh I've I've never
34:07
done well during the periods that tend to be the favorite for the discretionary crowd when
34:13
everything's just going straight up although I did I mean I I had my best
34:19
12-month period ever making 110 between uh July
34:25
2020 and July 2021 not a calendar year but but
34:31
that I consider a mediocre result because I mean I know of people who made over a thousand percent during that
34:37
period that was like the easiest time to trade since 1999. you know so okay I
34:42
caught a little bit of it um since July 2021 I've actually been
34:49
in a drawdown it's the longest drawdown I've ever had it hasn't been the deepest but um and I almost got back to Peak
34:57
equity um I don't remember exactly when it was sometime in um
35:04
mid early 22 maybe I don't know but um so
35:10
yeah I think it's it's been a trickier environment for sure since since July 2021 and I know a lot of other people
35:17
who have found that as well um that it's just harder to trade right now
35:22
yeah yeah and I guess you feel like uh trading is getting more difficult over time
35:28
yeah I mean it it seems whatever you're doing seems to kind of to stop working
35:34
at some point and but sometimes things start working again but um yeah it probably is I mean
35:42
hard to say you know it's it's very hard to sort out random fluctuations in in your results
35:48
versus some deeper change
35:54
yeah it's interesting because you know some Traders will say you know what the markets move the the same way you know
36:00
since you know the since that's been documented uh you know and I think there's a hint of Truth to that but at
36:06
the same time everyone's also competing to trying to find all of these edges so uh you know I was just curious on on
36:12
your thoughts there yeah I mean and if you get back far enough with bird's eye view you can see everything looks the
36:18
same or that's analogous to when people say uh the same patterns occur on every time frame
36:24
yeah maybe if you don't look at the details too closely but you know in the end you need you
36:30
need to have an entry Rule and an exit rule a knock at website solved too often and
36:35
that's what it comes down to I mean like Trend following is a great example it's a it's a classic strategy that everybody
36:41
knows and and it all comes down to whether you make enough on the outliers
36:46
to make up for all the chops that you pay and that comes down to the details of the strategy
36:52
yeah I was looking at some of the trend following uh you know Equity charts and I don't know if I'd have the mental
36:59
fortitude to be able to start with withstand some of the drawdowns that they yeah that they right well one trick
37:05
they use is to redefine drawdown to be only based on closed positions because if you you know if you if you
37:12
if you I mean if taking stocks taking an example uh you know Tesla if you bought
37:17
it in in mid 2020 and uh you know you've had a huge drawdown
37:22
but you're probably still up substantially yeah no that's a good point so it's it's actually uh could be
37:28
even uh crazier than they're telling us too so right right okay and then
37:34
um you know I guess you know one post that you made uh in Stock B um
37:39
that I thought was interesting uh you said that you know as someone who's traded mechanically for 21 years which
37:45
uh you know you're you're adding to that too and one of the only you know mechanical I think the only mechanical
37:51
uh Trader featured Market Wizards uh you said that you'll not make a killing you
37:56
know using a systematic approach your average return is around 20 which is still you know fantastic when you put it
38:03
up with you know New York professionals um and you've spent thousands of hour back testing and you uh you still feel
38:10
that the systematic side uh or systematic Traders are at a maybe a
38:15
disadvantage compared to discretionary Traders it depends what you're looking for I mean it's really hard to know
38:23
I mean certainly I mean Market Wizards will will highlight certain discretionary traders who have had long-term I mean one thing Jack is
38:30
always looking for is long-term consistency and and um and that's really the only thing that
38:36
got me into that book is that is a very long-term results that I have and and that's certainly an advantage of
38:42
systematic I think is that you you can do it longer um
38:48
you know I think the day-to-day especially short-term discretionary Trader it must wear you out after a
38:53
while to be having to watch the screen all day um but um
38:58
you know and I mean in both cases I think on the discretionary side it's
39:05
more true that the people we hear of our extreme outliers uh you know there's a lot of people who
39:11
try to trade and and just blow out an accountant and don't do it anymore and um you know so yeah I mean if you you
39:18
have a million people who try to do something you know you kind of have 0.1 percent who have extreme success and and
39:25
then those are the the kind of survivorship bias and I think that's less true in with mechanical trading or
39:32
systematic trading I mean it's there's there's not some some magic Edge that
39:39
people find I mean the the outlier there is of course uh Renaissance technology and and nobody knows exactly what they
39:45
do but from the reason excellent book about them um I can't remember the title of it but
39:51
um it um the man who solved the market is that is that yeah that
39:57
um you know although it's I mean his his ex what he really brought to the table was
40:04
was bring bringing together a bunch of brilliant people and managing them really well to you know
40:10
but uh anyway um yeah just getting all the all the
40:17
details right uh in a in your infrastructure getting your executions right there's just a lot of
40:23
details um right to work out yeah so there's less less room for area
40:29
I you know it's funny because um I don't really know if I have a choice of being discretionary or or
40:36
systematic I feel like it almost chooses you right so I have I've got a tech background and I I love programming
40:42
maybe just as much as trading um I I don't think I can stare at a
40:47
screen and and trade uh for for a long period yeah and another factor is what
40:52
kind of decision maker you are I I don't like to make decisions in general I
40:57
don't like to make quick decisions in particular um you know I like to get into the weeds
41:03
make sure I fully understand it and I go way too far in that direction yeah if that's your personality you're
41:10
not going to um right it's going to be harder to because discuss good discretionary trading is making lots of
41:17
decisions and effectively and often quickly amazing you know intuition which you
41:22
know after what how many hours of staring at a screen right and I'm sure it's there but um I can yeah and I don't
41:28
think I doubt there's any successful discretionary Traders you're just a purely intuitive or purely
41:35
situationally aware maybe some are I mean maybe the pit Traders were like that but but that's but they were all
41:41
very short-term they were really market makers so that's a different different game so um so if you had to say which you
41:49
know what's the superior Trader I guess that's a really silly question but um so you feel like that that you know
41:56
both systematic and discretionary obviously have their advantages and their disadvantages right
42:01
as with every other choice in training it comes down to finding what works for you like the the best system is the one
42:09
you can actually do I mean the best approach is the one you can actually do yeah it's funny that you mentioned that
42:15
because like as I was saying earlier I started from the the valuation perspective and while you know I did
42:22
pretty well at it it just you know over time uh you know I didn't like the drawdowns I wanted a you know faster
42:28
pace and um I like the maybe not I don't want to call it the game but it kind of is a
42:34
game but more of the challenge of of trading uh more than you know this valuation where you value something and
42:40
then you know hopefully in two to three years you're right right right yeah it's just not enough at bats
42:47
that's what that's what I couldn't tolerate about the idea of long-term investing either how do you become
42:52
statistically significant um yeah so I guess that's a great question
42:57
so when you're looking at your back tests um you know what are you looking at to determine statistical significance
43:03
you know how do you you know take a a new idea and then you know run that through to something that maybe you're
43:09
confident to trade with um yeah I don't I don't have a system
43:15
development system that's where I'm pretty intuitive when it comes to to
43:21
when I decide a system looks good enough but but I mean I want to see a back test
43:27
with let's say a thousand trades in it uh certainly at least 500
43:32
um maybe more and um
43:37
generally looking for I don't look at a whole lot of different stats um you know I look at the return and
43:44
then not just the maximum drawdown but I have a calculation of Maximum adverse Excursion which is sort of like
43:51
imagine you you were running a system and watching your net liquidation value in your IB account all day every day
43:58
what's the what's the worst number you would have seen um so which is a that's what Mae is good
44:05
for and that's kind of a good um you know psychological thing of
44:10
because I know from experience like when that number gets too bad it's it's uh it's hard to take
44:17
um so I'm you know I'm looking for I my usual threshold is I want to see
44:22
the the annual return be equal to or greater than the the draw
44:27
the max drawdown and and ideally the the Mae so kind of a one 1.0 Mar ratio as
44:34
people might say um and um and then I want to I also there's
44:42
some kind of qualitative aspects like I want to make sure I I understand something about the system it doesn't
44:48
have to be like a whole narrative but um you know I'll go through a lot of the trades and the back test and
44:55
just get a feel for the Dynamics or like I'll sort by biggest winners and then I'll sort by biggest losers and
45:01
understand what's going on there um you know maybe I'll I'll get an idea
45:07
for another filter to add but I don't go too far in that direction because you I mean in the end back testing is taking
45:13
the set of all possible trades that anyone could have made in the market and filtering them down to a profitable
45:18
subset we all know we all know the issue with that because they're because they all
45:24
happened in the past yeah exactly exactly and so when you're developing your strategies you're uh you
45:30
know Theory Theory driven versus you know data mining empirical we're kind of reversing reversing out that right so
45:36
yeah I did some experimenting with you know rudimentary data mining a decade or
45:43
two ago and um I don't know it just didn't resonate with me
45:48
you're trying to to forecast what's going to work in the future uh you know and and certainly I mean I
45:55
think you can pretty much rule out anything that has never worked in the past I mean I can't think of any reason to trade a strategy that has never
46:01
worked in the past um probably wanted to work more often than not if it had a big drawdown I want
46:07
to understand why I mean like I'm I I mean I trade meaner version shorts
46:13
which got absolutely killed in 2021 with the meme stocks
46:20
um in fact I have an example strategy in my software which which I put in there before that period uh
46:27
and and uh and it it had I mean it's its Equity curve looked like Bernie Madoff
46:32
but then it had an 80 drawdown in in January 2021 because it because it
46:37
bought gme I mean you know it's before that there
46:42
was never uh any any stock in the in the database that had that kind of a of a
46:48
gain uh you know because because nobody had discovered options gamma squeeze yet
46:54
um I guess so yeah you never know what's what might happen
47:00
yeah somehow I mean I that's still like I could have that could have happened to
47:06
me somehow just by random luck my my the particular version I was running didn't trade it and I've I've since taken some
47:14
I do some things in my execution software now that um that I don't back test I I just consider
47:21
them as like Risk Management and one of them is it occurred to me during the meme stock squeeze that uh
47:27
you know ID has changes the margin requirements for individual stocks on you know every day
47:34
or actually multiple times a day um so just set a cap stay you know I figure if I if IB says
47:41
says they need 100 margin for this thing uh maybe I just want short
47:47
so I could do that that saved me from some of the bigger losers because I put
47:52
that in as soon as I saw that thing starting to happen Okay so perfect I think we're we're getting pretty close
47:57
to to an hour here um you know kind of wanted to talk about um now is there any uh
48:05
books or podcasts or anything like that or any uh piece of advice that you would
48:11
like to share with the Traders you know new systematic Traders or even discretionary Traders for that matter of of how to you know get started and and
48:19
maybe have uh half the half the great track record that you do
48:24
um I mean the main advice I would have for for anybody is to get your hands dirty and do your own
48:31
experimenting you know you've got to you've got to get even if you're just discretionary you know you've got to do some sort of back
48:37
testing and I mean and if you're going to do sort of visual
48:42
um manual back testing on charts like a lot of discretionaries of traders who be careful because you'll confirmation bias
48:49
is way too easy like if you're looking for a nice looking breakout setups like they do on stock being just looking for all the winners
48:55
you're not checking uh you really need to say okay this setup
49:01
happened uh it's harder to spot the ones where the setup happened and then failed
49:06
because it just looks like noise on the on the chart so that's where back testing can be very
49:12
helpful but but the the key advice is to set up an infrastructure that lets you
49:18
do your own experimenting do you you know just constantly
49:23
iterating with what if I try this how would that have done what if I try that how would that have done and so on
49:29
um yeah there are I can't point to to a specific book that
49:36
tells you exactly how to do it uh you just you really have to try I mean the best thing is to join a um
49:43
join a community and of other people who are doing the same thing and um I'm happy to now have built a
49:50
really nice community in the real test user Forum which I keep closed uh you know except to people who actually use
49:57
the software um and um and that creates an environment where people are comfortable
50:02
sharing ideas and um discussing what they're what they're trying
50:09
so that I mean a podcast um you know I listen to all the usual I
50:15
certainly listen to all the ones I've ever been on uh and oh and some some others um even though I'm not into Trend
50:22
falling I really like the top Traders unplugged series that Neil's Kester Parson has
50:28
um I listened to that one too yeah yeah and I you know I used to listen to all the Andrew swans got better system
50:34
Trader one so he hasn't doesn't need to do them anymore the chat with Traders I listen to most
50:40
of them although most of that had more of a discretionary skew but it's going in a New Direction now that's that
50:45
that's interesting ahead I was delighted this last one they had Patrick Boyle on who's who's uh different sort of guy but
50:53
I love his podcast series called on finance Marie talks about uh you know there's a lot of
50:59
Market history in there and um and so on and I had no idea about his
51:05
background until I listened to his recent chat with Traders episode in it uh yeah he has quite a trading background
51:12
he used to work for Victor leaderhoffer and he he discovered early on that he was a
51:18
systematic Trader himself and started corresponding with Victor and got
51:23
offered a job okay that's great that's great perfect okay and then uh yeah can you
51:31
share us a little bit uh more about real tests for for those that that might be interested in uh you know using your
51:37
software yeah yeah I mean I I have a little rudimentary website called
51:43
mhptrading.com and and you can read about it there and download it download
51:49
it and try it um it's um
51:54
it's uh you know back testing software you have to you have to import data you
51:59
can I recommend nor gate but you can also use Yahoo to get started with some free data
52:05
um or your own CSV files if you have them or metastop data if you have it
52:10
um it comes with a lot of example strategies and Scripts that help you get started and um
52:18
it requires a little bit of scripting but I've had quite a few people who have tried to learn other
52:24
practicing software that that requires more like a programming language um
52:29
and couldn't get anywhere and the real test evolves with sort of like it's more like filling out a form in the the older
52:36
versions you would put in a like you would put in your entry setup formula which is sort of it's more like a scan
52:42
criteria that you would write um and and so that's still how it works uh except it's you rather than filling
52:50
out a form you type the label on the script like you'll type entry setup and then you'll put in your entry setup formula and you'll type exit Rule and
52:55
you put in an exit rule formula and so on um and and these formulas are grouped
53:02
into named strategies so you can put as many strategies as you want and run the tested you get
53:08
the uh the stats for all of them combined and you can see all the equity curves separately as well as the
53:14
combined curve not and so on so yeah the goal is to my goal with it
53:22
is to have to make a tool that's really easy to interact with
53:27
and really fast so you can kind of get into that real-time experimental
53:32
experimentation Loop that I was talking about earlier um because that's how you learn it
53:39
so I did sign up for for the beta a long time ago uh to check it out and I I felt
53:45
that um it was actually very easy to use because right you know when I'm I'm you I use Python for for my trading
53:51
strategies and uh you know you don't have to necessarily know a programming language you just have your you know
53:58
your essentially scripts which if I recall correctly I mean just very simple
54:03
um so you know for those that uh you know might not want a programming language like uh python or rust or C or
54:10
whatever that just wants something that's powerful I think you should probably check uh check out Marston's real test well thanks for the plug yeah
54:16
as a as a sort of um indicator of that my eight strategies that I'm running
54:22
right now the the entire script which they're all in including all the indicator calculations that they use in
54:28
the the part that says how to import the data and all the settings and everything is under 200 lines long
54:34
yeah that's that's phenomenal and it's quick right I know that uh yeah unfortunately Python's easy to to
54:41
program in but I'm actually with some of my strategies you know I'll I'll run it and then you know go have dinner and
54:47
then have to come back down later and I'm like oh this is starting to be a little bit unwieldy but yeah right I
54:53
mean a simple strategy like if you wanted to test its classic NASDAQ 100 rotational strategy going back to 1990
55:01
using the historical index constituents I mean it takes I don't know two or three seconds to run that back says
55:09
um yeah real test is my my when I was working as a software engineer is
55:14
primarily a Windows C programmer using the original Windows API you know just
55:19
like Microsoft used to build their original office suite and so on um and
55:24
For Better or For Worse real test is still written in that language which makes it harder and harder to find
55:31
anyone else who could ever maintain it but um it uh it it's yeah it's small
55:37
fast it's it's completely built on the original Windows API requires windows I
55:42
mean the good news is a lot of people run it on a Mac using whatever they call those things parallels or yeah but but
55:49
yeah you do have to have a Windows installation for it okay perfect so then if you're interested in you know algorithmic
55:56
trading and multi-strategy something easier than a programming language uh check it out uh so Marston I I really
56:02
appreciate your time today as I mentioned earlier you know essentially what am I trading Heroes pretty cool to
56:08
have you as the first podcast guest very exciting uh because quite frankly there's you know really not uh that many
56:15
systematic Traders out there especially uh that are that are Market Wizards I'd like to thank you once again and uh you
56:21
know like to stay in touch thank you so much I'm sure we will yeah okay good luck with this series I'm sure you'll do
56:26
well with it and it's been really great to meet you Perfect all right thanks person take care bye you've just
56:32
listened to the first ever Alpha cast episode if you like this conversation with Marston Parker the legendary
56:38
systematic Trader please give it a five star rating or thumbs up depending upon what platform you're on so that way
56:44
other Traders and investors can be lucky enough to listen to this fantastic guest too thank you
Investing
Stocks & Bonds
People & Society
Brokerages & Day Trading