Up next in 10
Uncover the remarkable strategies of a 22-year-old trading whiz who's beating the market! In this eye-opening episode of the Analyzing Alpha Podcast, we sit down with Goshawk Trades, a young master of algorithmic futures and small-cap stock trading. Whether you're a beginner or a seasoned trader, this educational and informational interview offers invaluable insights to elevate your trading game.
āļø Learn the secrets of legendary traders and investors:
https://bit.ly/4aCna7i
š Follow along:
https://analyzingalpha.com/episode-007-goshawk-trades
In this interview, Goshawk covers:
- How he got started with algo trading and his journey so far at age 22
- The specifics of several of his profitable futures trading strategies
- Tips for robust backtesting and avoiding overfitting
- Managing emotions and psychology with algorithmic systems
- Advice for new traders looking to get into algo trading
- How to balance a job while learning algo trading
- His favorite learning resources and books for algo trading
ā±ļø Video Chapters
0:00 Introduction
Show More Show Less View Video Transcript
0:00
hey manir I'm so excited to have you today uh can't wait to discuss uh all the algo trading uh with you and how you
0:06
got into it so with that being said how did you get into this crazy trading game yeah so primarily uh when I was quite
0:12
young I was always interested in finance and kind of markets um originally I started with dividend investing which
0:17
didn't make me much money um and then kind of from there I got into the classic um training kind of routs of training discretionary so a lot of the
0:24
beginning it was uh Lodge caps Tesla Apple using kind of resistance lines things like that um and then I've been
0:29
trading for about four years and a half now and got into more algo trading for the past two and a half uh give or take
0:35
um that was mainly to diversify my trading um from not just being you know discretionally trading and using kind of
0:40
manual execution and then taking advantages of Futures where you can you know be very algorithmic and it can be really um kind of profitable and I know
0:46
you said when you were younger if you don't mind sharing how old are you because I'm pretty sure you're you're getting a a much headed start on on me
0:53
right uh currently I'm 22 and uh yeah I guess it's starting off quite young fantastic and you started quoting when
0:59
you were 12 yeah so um primarily I started coding for Minecraft servers I'm just kind of developed my own kind of
1:04
stuff um that was around 10 10 12 um and then more getting focused into pyone um
1:09
a bit more focused when I was doing some statistics in my kind of degree um and then trading wise really I just focused
1:15
for the past like three years um mainly on trading kind of code and then I guess the question is um you know why did you
1:21
get into you obviously you have the tech background which is I I love Tech and you know I started trading discretionary
1:27
but I actually missed uh the programming and all that stuff but why did you get into uh trading algorithmically yeah so
1:34
I kind of touched on it just a bit before it was saying uh mainly to diversify so primarily my main Market
1:40
where I started or at least started to see consistency with small caps um and small caps is great you have huge
1:45
volatility um you there's a very obvious Edge statistically so it's really easy on data analysis kind of side of things
1:51
um and a lot of time you're kind of just filtering to kind of improve that edge or reduce your risk um but the other
1:56
side of smoke apps is there's you know there's halts there T12 HT uh for people that don't know that's kind of just a news halt so you could
2:02
get stuck in that for a day like a month for a year um and sometimes during that time you can even blow up just by having
2:08
to pay for locates which is a whole another side of small caps where if you're going short uh you're going to need to pay for locates and depending on
2:14
Brokers those can be very expensive so there's a lot of uh inefficiencies you can exploit but there's also like a lot of risk um to it as well um so the main
2:21
reason I started getting to Futures was primarily first off was The Leverage is pretty much insane uh so you don't need
2:28
that much Capital to start kind of toying around uh with Futures uh like even with a brokerage side of things you
2:33
can get into ninja Trader I think like the minimum deposit is like 500 or like a grand um so that's kind of how it
2:39
mainly started um and then from there I started to trade it a bit discretionary to get some ideas um and then kind of
2:45
from there get into the algo side of things where I want the Futures to be fully kind of hands off I'll check it in the morning I'll check it in the evening
2:51
make sure there's no like bugs have alerts things like that um but then focus on small caps and kind of diversify my portfolio to an extent
2:57
where I'm technically trading two markets at the same time you mentioned ninja Trader so did you roll your own
3:03
platform did you are you trading with Ninja Trader right now I'm curious on that because um I ended up creating my
3:09
own trading platform and I think it you know at least it might have been a mistake I'm just I'm just curious on uh
3:15
the direction you R for uh trading execution things like that you know hats off to you for creating your own um I
3:21
looked into it and it was very daunting so I was like nope I think I'm just going to stick with simple here um and I
3:26
just kind of went with Ninja Trader um there's definitely do a lot of good pros of ninja Trader um to name a few um the
3:33
back testing optimization walk forward testing all of those are pretty good um there are some nuances here and there
3:38
that I don't particularly like of ninja Trader like some of their calculations for performance metrics are are not very
3:43
good um so you can't have to export some of that trade data and do your own calculations um but apart from that at
3:48
least on the execution side of things they're they're pretty much perfect I've never really had an issue with them um
3:53
and they're very easy to deploy on like vpss or things like that um compared to you know moving over your kind of
3:59
platform something like that sure sure and can you do like multiple time frame analysis and things like that on on
4:05
ninja Trader and those types of platforms yeah you can do a decent amount the only thing you can't do is you you can't do optimizations of
4:11
multiple time frames all at once um with those strategies um I think another thing that most people wouldn't like is
4:16
you can't do portfolio optimization or like portfolio analysis in N Trader at least not my knowledge um so a lot of
4:23
times I have to export all the trade data into Excel or and then like kind of use python or Excel to kind of do that
4:28
analysis um so it's definitely it to some extent it's rudimentary um but I
4:33
I've never found it to be hindrance to at least getting you know some profitable strategies and at the moment I only run around four to five give or
4:40
take um so maybe once I'm getting to the 30s and 40s then I I'll probably try and
4:45
move off it um but I've even seen for some people like um Jacob Al Melly I think is how you say say his second name
4:52
he's done very well with Ninja Trader and he runs apparently around 30 strategies and for those of you that are curious I would actually recommend going
4:58
that route and not uh I ran into a lot of challenges rolling my own and now I'm uh going through another iteration so uh
5:06
so yeah unless you really really love coding don't roll your up just just so you know um okay perfect at least that's
5:11
my my opinion yeah it's a big challenge yeah it definitely was there's just so
5:17
many things that you don't think of that like yeah it was it's it was a challenge and it still is but anyways uh can you
5:23
describe any of those for trading uh strategies obviously if you don't want to you don't have to get into any of the specific details but just curious how
5:29
you're currently set up yeah sure um so I'll I'll go over kind of like a couple of the main ones so the first of a main
5:35
one that I saw good kind of consistency and success with was a trend line strategy uh main reason for it was when
5:41
I was kind of getting into futures um for for a time I just traded it discretionally um for like three to four
5:46
months just on like propop accounts and just on Sim um just to kind of help idea flow just to see you know what stuck out
5:52
to me what I think could work um and one of the first things I kind of saw myself uh start using was mainly trend lines um
5:58
so I started to to kind of try and back test them um that was a whole ordal in itself trying to quantify trend lines
6:03
isn't as easy as as you think it may be um so I kind of went into that route and then that strategy is proven uh very
6:09
profitable over the time pretty much what it is is I'm just using trend lines and then the slope of that trend line uh
6:14
to assess whether it's a short or long trade with some kind of sentiment filters as well based on like higher time frames um and I'm just using atrs
6:21
really uh for exits and profits so quite simple uh in its kind of usability I think the only complex part of it is the
6:27
the trend line calculation um the next was a machine learning scalping um strategy so that's on renco bars it's
6:33
it's custom renco bars if you go with the original your back tests are all going to be floored so make sure to have some kind of custom ones that have the
6:39
Wicks and everything um and that one to be honest it's just doing meme reversion around a fair price and it's really just
6:45
kind of calculating a fair price using time uh and a couple other kind of filters um that one's definitely a bit
6:50
more complex I've had issues with potentially thinking it's being like over fits especially because the whole machine learning side of things um still
6:57
being profitable but it's taken a lot lot of tweaking um and a lot of kind of testing and being worried that I'm
7:03
overfitting um so that one probably taught me a good amount uh and then another one I'll mention is a very easy
7:08
one that I think so many people Overlook is just Mas uh so it's just an MA slope strategy um it's got a couple other
7:14
filters here and there um to kind of just help with sentiment but realistically it is just a slope
7:19
indicator and an ma uh indicator and it just helps uh with doing a kind of trend strategy along those lines okay no that
7:26
that's fantastic and how did you come up with those ideas you know I guess I'll just be completely honest most of the
7:31
stuff that I've come up with has been stolen right I find traders who I uh you
7:37
know really enjoy and obviously learn from them I've got about one alpha that I came up with by myself but I'm just
7:43
curious on on your thoughts there I think that I think if you hear someone say that that they're they're truly an algorithmic Trader I would say pretty
7:48
much the same um I don't think any of those are really original I've taken Parts pretty much from everyone um
7:54
normally my kind of resources to go to is either podcast like yourself or other people's podcasts um really Twitter I
8:00
think Twitter is an amazing resource so I go through Twitter a lot um and then lastly kind of books um books have been kind of a Hidden Gem for me um I never
8:07
really used to read them that intensively but at the moment I try and at least go through a book once in a while um because you can find some
8:14
really good hidden kind of old strategies in there um most of the time I will say at least in old books they
8:19
won't work straight off you'll have to kind of understand the idea that they're trying to go for and then apply it on to
8:25
kind of recent markets just because volatility and things like that have changed so much since some of these books have kind of released um some of
8:31
them still work but that's kind of my main way of doing it do you have any favorites by any chance any favorite
8:36
books yeah sure uh I think I did a tweet on this recently and it's definitely still my favorite book is a evidence-based technical analysis by
8:43
David awon I think is how you say the second name um another one I love is trade like a hedge fund that one's
8:49
really good um main reason I like that one is some of the strategies are quite good that's where I started testing a
8:55
Pez trading strategy uh unilateral sorry Pez trading strategy um kind of basically stole the code from that book
9:00
and applied a couple different things to it um I don't have it currently running at the moment I I did run it for a
9:06
decent amount of time it did okay I think the main issue I had with that was having conviction in it because at least
9:12
for pair's trading strategies they're quite hard to back test historically uh when you're trying to do like NQ Andes
9:18
or something like that on ninja Trader um so I I kind of stayed away from it just because I couldn't test it properly
9:23
I did a lot of forward testing of it which I'm you know it's good but I'm not a huge fan of only doing forward testing
9:29
um so I ran it for a bit and just kind of left it um what's another good book I think Robert C's 30 future strategies
9:37
book is quite good um some of the points he makes I don't fully agree with but a lot of the strategies can be quite
9:42
useful to kind of take ideas from and apply and you mentioned um forward testing so um you know I've got um a
9:49
fair amount of discretionary Traders and algic Traders here can you sort of just kind of back up a little bit and talk
9:55
about your back testing process and maybe what a good uh strategy uh you know what a what a good testing
10:01
pipeline looks like at least from your perspective yeah sure um so I'll say like first off that none of this is
10:06
original I've pretty much taken this from a lot of the the the goats as You' say in quantitative trading Ernest Chan
10:12
a lot of those other people um but primarily starts with uh with idea generation as we kind of touched on I
10:17
think that's one of the most crucial parts to be honest I think a lot of the time you should probably be spending most of your time in idea generation and
10:23
then quickly testing and seeing you know what works um I think that's normally the way to go the kind of mistake I made
10:29
before was spending way too much time doing not uh just doing iterating on
10:35
back testing um most of the time if you find a good strategy you'll sort of know instantly if you do a ton of iterating
10:40
then you're just going down a slippery slope uh but at least for the pipeline yeah so it starts with uh idea generation then I just go straight into
10:46
quantification and testing so normally that's if the idea isn't you know clearly laid out I'll have to apply some
10:52
assumptions of how the calculating it um then go into back testing either with python or ninja Trader if it's through
10:58
ninja Trader then I'm coding in cop and if it's for equities I'm using python like polygon data and there's a couple
11:04
of other apis you can use but at least for me polygon being a great resource um then after that I'll go into robustness
11:10
testing so that's you know multic parameter sensitivity testing out of sample testing um those are only the
11:16
main free I kind of hit um I have templates I just kind of use in in my own python code to calculate those um
11:22
and then I do full testing if everything looks good off the there uh then I'll full test it in a Sim account or a very small Live account for
11:29
depends on the strategy and the amount of frequency I can get through it but realistically I'll probably test it for about a month two months um if it's a
11:36
super high frequency strategy then I'll maybe only test it for a week or two uh but most of the time most strategies I'm
11:41
kind of using are doing about one trade a day maybe two trades a day um and then I'll take it fully live if everything's
11:46
looking good off there perfect no that's very helpful and um just so um the audience understands can you explain uh
11:52
you know Monte Carlo and uh forward testing yeah sure uh so Monte Carlo um there's kind of two main methods of it
11:59
you've got distribution uh redistribution sorry so let's say you have uh 200 trades in your back test um
12:04
it's then picking those trades at random and putting them in a random order um and then you've got kind of resampling which is where it'll take any of those
12:11
200 trades and it could even duplicate them so you know it could pick your top five worst losses if for example and see
12:17
how your strategy would Prevail and then it runs that simulation you know whatever amount you want it to so
12:22
normally I do about 300 to like a thousand uh and then it plots the equity curve and you can just kind of see the
12:27
general statistics of all those uh Equity curves um the kind of main things you want from a Monte Carlo to look at
12:33
is first obviously the average performance is over a thousand of those simulations you know is it still profitable is it still you trending
12:39
upwards uh and another one that's really good is just for the risk analysis side of things so seeing what was the Max draw down on the worst outcome um you
12:47
know there's some rules of fums you can do so most of the time I'm taking that one and that's how I'm basing my kind of
12:52
My Equity that I need to deploy to that strategy or the amount of risk management um other people Tans are by
12:57
two to be extra safe um you can really kind of go whatever way you want there but it's a good thing you should at least look at uh and then for forward
13:03
testing it's it's pretty simple you're just kind of turning it into an algo um and you're running it on a Sim account
13:09
and um kind of like just just testing it over live data um and you're kind of comparing the performance to your back
13:15
test performance um different ways to do this if it's a super like Market cycle
13:20
kind of strategy um it depends on the market cycle then you'll probably want to compare it to when it was in a similar Market cycle um if it isn't then
13:26
you can just compare it to the last kind of month of performance in your back test um and see how it's kind of fairing up using performance metrics the equity
13:33
curve um those are normally the ways I'm kind of testing it and you know for those of you that aren't algorithmic I
13:39
mean um these are still things that you could potentially be doing or you know paying someone else uh to do for you
13:46
right to give you confidence and you help uh discretionary Traders uh understand their uh their strategies
13:52
yeah so primarily I run a small kind of business called on bias trading um primarily it's just a coding service for
13:58
Traders um one thing I've got a lot of questions from people in my DMs or just people messaging me overall is hey uh I
14:05
need this coded and at the beginning I was just kind of doing favors I was doing like small side things here uh but
14:10
that's just kind of grown to where now I'm I'm working with Traders either develop their strategies into full alos or we're doing full you know um kind of
14:17
full proper back tests um over five 10 years of data using whatever Source we need to for that strategy um so most of
14:24
the time people come with me either with an idea of the strategy or like a full criteria list so you know exits and
14:29
everything and then we either turn it into an algo or we do back testing depending what the kind of interested them and then I guess from the discretionary point of view have you
14:35
noticed any like common themes or or challenges that maybe um you know those discretionary Traders are are running
14:42
into or overlooking um do you mean just like primarily if you're just trading discretionary or if you're trying to get
14:48
into systematic trading yeah yeah I I left it I purposely left it broad because I don't want to like pigeon hole
14:54
I'm just curious like um you know obviously you know discretionary Traders have a lot of like intuitive
14:59
understanding of the markets and um you know experience but is there something that like a common theme that
15:06
discretionary traders who aren't algorithmic often uh mistake or take for granted or or a specific area that they
15:12
could improve yeah sure uh I think one that comes to mind straight away um based on recent kind of price action for
15:18
small caps is people are insanely recently bias um just like an incredible amount when you're really discretionary
15:25
um you'll have like a week of losses or you'll have a bad streak for a week or two weeks and then instantly I see
15:30
everyone on Twitter go oh I need to change my strategy oh this isn't working anymore I got to change my entry
15:36
criteria and all that kind of things like that where realistically I never heard really of a strategy that hasn't lost for two weeks at some point
15:43
especially if you're in like a very cyclical Market such as small caps um you're going to have times where it goes against you and you're going to have to
15:48
handle that that's kind of the whole point um of trying to have that strategy now don't get me wrong it's tough It's it's not a fun experience um but that's
15:55
just something that kind of comes with the game that's great insight and then in regards to some challenges for
16:00
algorithmic trading uh you know maybe uh can you shed some light on some of the challenges that you faced um you know
16:06
with algorithmic trading or some of the misunderstandings um that maybe a discretionary person who is looking to
16:12
get into algic trading might face yeah sure um I think the first one I I touched on a bit before was a lot of
16:18
people get into systematic training including myself um and you'll you'll go down the back testing route and you'll
16:24
start twinker with all your different parameters trying to get the best profit curve trying to get those draw Downs the
16:30
lowest um because I think it's really easy to assume in the beginning that if you get that back test showing really
16:35
good results then boom you're you're a profitable Trader now you just have to take that live and you'll keep that performance um and it's very unintuitive
16:42
to realize that technically the more you optimize this the less likely you're actually going to do in profits
16:48
afterwards um so most of the time I've kind of taken back and the back testing
16:53
process is still crucial don't get me wrong it's just I don't spend as much time twinker with those kind of parameters and trying to get the the
16:58
best ones I'm I'm normally doing a very light optimization um and just kind of taking normally the the third top result
17:05
instead of the first um things like that just to help with the overfitting um the annoying thing is that there is no you
17:11
know ultimate solution to avoid overfitting you you got to do good robustness testing you have to have uh
17:17
learn from your mistakes of when you've overfitted the strategy and have some sort of common sense with it um and that kind of comes into the part of being a
17:23
quantitative Trader or trying to get into algorithmic trading um another one I I see quite often or at least on you
17:29
know when people are trying to advertise their alos or something like that um is that it removes all emotion you're going to have no emotion when running the
17:35
strategy um it's just it's not true yeah i' like to say I think this definitely isn't from me I've taken this from
17:41
someone but it's someone was basically saying that it takes the emotions out of individual trades but it places them on
17:46
the strategy um on you know a day-to-day basis you probably won't feel much of emotion but if you still suffer a week
17:52
of losses you're going to be thinking that algo doesn't work and you need to change it and you need to turn it off um and all those kind of things still come
17:58
into play so a lot of it is still emotions it's just a different degree of them and at least for me handling um
18:06
more of a long-term amount of emotions compared to an intraday emotion type is way easier for me to handle so I can you
18:13
know take the time to look over those week's performance and be like is this realistic do I need to change it that's a lot easier for me to process compared
18:19
to um training discretionally and getting emotional on that one trade that I I should have held differently or exited differently as well sure no I
18:25
agree with that the uh the emotions and discretion trading i' I felt the recency bias I don't know if you're trading
18:32
discre if you can't um but uh but yeah that's still one of those things like I
18:37
still even with my OS I mean it is it's on this strategy and I've you know uh turned them off and turned them on
18:43
probably when I shouldn't but anyways um you know we're all subject to those uh those emotional uh challenges um in
18:51
regards to maybe some of the biggest risks what do you think the biggest risks uh for for algorithmic trading is
18:57
and and how do you manage it um yeah probably the first ones to say are probably code errors or platform errors
19:04
um that's why I think it is really vital you're picking a reliable platform to kind of run your AOS um at least for me
19:09
that's why I've liked ninja grader I've never really ever had any issues with them um most of the time I'll say
19:15
actually most issues come during a data source issue um so if you're using kinetic or you know another data source
19:21
in ninja Trader or if you're on equities you know using like an API um sometimes there can be data issues there can be
19:27
discrepancies and those can cause a lot of weird bugs to happen or weird orders or executions um so
19:33
having I will say that it's it's going to happen at some point um I don't think there's like any crucial way you can
19:38
avoid it it just comes with a game with being in algorithmic trading same with risks come with you know day trading it's just a kind of part of the game um
19:46
but the best ways I've tried to avoid it is making sure you have Broker safety measures in place so making sure you have that Max loss those max amount of
19:52
entries max amount of orders um those kind of things to make sure um at some level your broker is kind of responsible
19:59
to make sure that you you know those those are held in place um and at least that can maybe help mitigate some of the
20:04
damage um another one is making sure to use checklists uh make sure you have alerts on your phone make sure you you
20:10
know when you're in a position when you're out of a position all those kind of things or when an order is sent um those can be finicky to set up sometimes
20:16
but depending on your platform all that kind of stuff you normally can do it um and then lastly I'd say always use small
20:22
size um obviously small size is relative though so as you're getting into a bigger account you can start using more
20:27
money and the profits can grow um but I've always tried to be a bit more risk averse to some extent um to make sure
20:33
that if there was this terrible circumstance I I don't lose 100% of my account you know even if I lose 50
20:38
that's a lot better outcome than being fully wiped out if that makes sense no no it definitely does And in regards to
20:44
the strategy uh challenges how do you go about I guess understanding both the uh
20:51
you know the challenges that that it may face and how do you mitigate that uh I think this is a tough question this is
20:56
like the the magical question if you if you you know when to turn off your strategies then you're you're in a really good place um for for me at least
21:03
what I've seen to work is is really actually not applying much um a lot of people the the classic advice is to put
21:09
an MA on your Equity curve and you know turn it off when it goes below and turn it back on when it goes above um the the
21:14
big issue with that one that I at least hopefully I'll help someone with is depending on what EMA length or ma
21:21
length you choose you're either going to be way too early or you're going to be way too late um and I don't know if there is that magical ma length that is
21:28
is going to solve all your issues at least I haven't found it um so most of the time I I use it as a measure to help
21:34
me understand maybe if I should you know lower my risk lower my size I think that sometimes can be quite helpful um I try
21:40
to use more of a longer term Ma on My Equity curve to kind of help with that um but a part of those things it really
21:46
comes down to like proving to your psyche that you you have a good system and you need to be confident with it so I try and immerse myself with as much
21:54
kind of data information as I can um do as many tests as uh Poss make sure you're doing good back testing spending
21:59
a good amount of time with it I think there's some sort of um human bias or quote that the more time you spend in
22:05
something the more uh confidence and you know ability you'll have to withstand with it now that can also bite you in
22:11
the butt um if you spend you know 100 hours on a strategy that isn't profitable you're still going to think it's profitable um or you're still going
22:16
to try and make it profitable um but I try and apply that kind of methology to make sure I stick to systems now I still
22:22
make mistakes I probably still have turned off systems where I shouldn't have um it just it's a Lear process I
22:28
think uh I maybe in 5 years I'll have kind of a better answer for that but at least at the moment I don't think there's any like perfect solution yeah
22:35
no I think you're you're absolutely right I don't think there's a perfect solution I think the solution isn't to do what I did when I first started I as
22:41
I came up with this huge strategy you know you think oh I'm going to come with the Holy Grail and and you know it got
22:46
to so complex to at one point that I couldn't follow it and then um what eventually happened is I ended up
22:52
creating a portfolio of strategies and I think that when you have multiple strategies it helps take the focus off
22:58
of one individual strategy and you're not like stressed out all the time so so yeah I was nowo Trader and I was stressed out of my mind but yeah having
23:06
multiple strategies I think is a is also a way to diversify and uh and having that so so know those are all great
23:12
great ideas so cool okay then how about um did you have did you suffer any um
23:19
you know big setbacks during your your trading Journey yeah I actually I also want to touch on your portfolio mention
23:25
um is a is a really good thing to try and kind of take away from from the Holy Grail of one strategy and kind of use R's mythology of holy grail's you know
23:32
uncorrelated strategies and if you go that route you're you're normally going to go the better way of it um and you
23:37
can also just focus on it it takes away that one strategy has to be your you know 500k a year profit or something you
23:44
can you know choose a lot of small strategies that add up to a good amount of profit and reduce your kind of draw Downs um that's normally the way I'm I'm
23:51
trying to go so obviously only at the moment I only have four strategies so I'm definitely not the best with it um but as you add more and kind of test
23:57
more then that's kind of the way you want to go and stop taking out of focus of spending so much time on just one strategy as you mentioned um but for a
24:04
significant setback I think the one that comes to mind the easiest um I did a tweet about this ages ago I I think it
24:10
was I lost about 1.7 1.8k in a day simply just because of a VPS issue um
24:15
pretty much what happened was uh I was trading um my algas were trading on ninja Trader VPS and it lost connection
24:22
for a bit and it was the second very unlucky um second is when it was placing orders it didn't place a stop order and
24:28
for some reason it doubled my order still don't know really why that why that happened to this day um and then by the time I checked I was down 1.7k on
24:36
the day and I think at the time my risk was probably about like $700 so that was a crippling loss for me at least at the
24:42
time um obviously it it didn't blow up my account or anything like that but it it was a very shitty day um to kind of
24:47
deal with um main ways I kind of tried to bounce back from there was I just introduced a lot more safety procedures
24:53
I made sure I had alerts on my phone if an order was being placed or you know um being executed uh made sure to have some
24:59
kind of VPS backups to some extent so I kind of had like a a just a backup server in place just to handle any of
25:06
those orders um and then really alerts I I think those are really the best ways you can do it um at the time I think my
25:11
Max loss was about 1.7k so it pretty much hit it um for that day um and
25:16
that's normally the way you go through it but I think even I I was talking with another Trader recently and he had a
25:22
very similar situation um and as we kind of mentioned before it just kind of comes with the game sometimes you're going to get very unlucky um sometimes
25:28
you'll be stuck in a hole and you'll have no idea why um and it's the same thing kind of algorithmic trading no
25:33
that's absolutely true I mean algorithmic trading alleviates some of the some of the challenges but it introduces a whole host of other ones um
25:40
you know obviously I'm still biased and I think algorithmic trading at least for me is a way to go but again uh you know
25:46
more challenges with the uh you know come become introduced and then uh what advice would you give to someone you
25:53
know starting algorithmic trading um you know would you suggest someone just jumps right into it or do you suggest
25:59
they start discretionary what's your thoughts there I've given this a decent amount of thought I think to be good at
26:05
algorithmic trading you you do need to trade discretionally for some amount of time um I think it's also why um someone
26:11
says like phds in in algo or at least in in trading don't really mix that well you kind of need some trading experience
26:17
um a lot of the time if you're just developing strategies and you have no idea how it feels like to trade those strategies you're going to kind of be in
26:23
for a rude awakening once you you know set those live um and a lot of things like even when I started trading there's
26:29
a lot of um mistakes or ignorance that I didn't even know I had um sometimes just
26:34
understanding slippage um on different markets and how that can affect it um different Holts and things like that can apply to different markets um so I think
26:41
to some extent you do need some trading discretionary uh like experience um doesn't have to be too much I'd say like
26:46
I think if you probably traded if if we're saying like from Ground Zero if you probably traded for six months to a year um on a Sim account or just you
26:53
know watching a lot of content you could probably then start getting into algorithmic trading and start to code based on your coding ability um for more
26:59
of the coding side of things or at least algorithmic trading I would say spend a lot of time experimenting um don't stay
27:06
too much on one idea stay um just a lot on your curiosity of testing a lot of ideas instead of just focusing on one um
27:14
I'll try and assure you that what one idea is not going to be a holy girl as we said um you're not going to find that
27:19
one strategy that's going to make you a millionaire it's probably going to be a combination of a lot so you need to spend that time experimenting and see
27:25
what kind of works um another piece on that is you don't need to reinvent the wheel um for when I was getting into
27:31
Futures Trading I really thought you know I've been successful in small caps I'm going to be able to find this amazing Edge in Futures I don't need to
27:38
learn from other people or at least that's you know some unconscious way of how I was thinking about it um and what I learned quickly was no I don't have
27:44
any really good ideas um and I need to start looking at what other people are doing and and try and apply it for myself and adapt from there and I think
27:51
that's normally the the best way to go for it no that that's fantastic and I 100% agree I don't I think you have to
27:57
trade uh discretionary um at least for me I know it might sound strange but it's part of my idea Generation Um and
28:05
there's certain nuances when you're trading yourself that you don't understand when you're at uh a higher level so when people say oh I'm just
28:11
100% algorithmic if they don't have you know actual trading experience um I feel
28:18
like there's there's a lot lost and I know you said you still do some discretionary trading right do you feel the same same way like you're still like
28:25
catching these little nuances when you're Trading uh that way yeah so I I will say my discre of trading in small caps um helps
28:32
me I guess keep on the ball with trading and those kind of nuances obviously it's not trading Futures markets um I do I do
28:38
look over the Futures markets kind of every day and see what happened um see where I thought would be a good trade kind of just helping my idea Generation
28:46
Um it's kind of a whole different topic of why I trade still discretionary um or at least have that portion of discretionary trading in small caps it's
28:52
mainly because there are some things and I think most algorithmic Traders will learn this at some point is that there
28:57
are nuances which sometimes experience really can beat an algo on um so just for understanding Market Market Cycles
29:04
to some extent and small caps can be quite a complicated weighing uh issue um you've got lots of different parameters you can look at all of them kind of do
29:10
hold significance to some extent uh but based on the day and the time um those
29:16
can be really different in weighing sometimes the news is going to matter way more and the other times it's going to be because of how much volume it's
29:21
just had in the morning and the float rotation um so those are the issues that sometimes experience can beat mic
29:27
trading on now I will say if you don't have that experience algorithmic trading will always beat it uh thinking in a kind of quantification way is always
29:34
going to be better than if you're just trying to just trade it discretionally with no experience um but yeah no I
29:39
think you do need to have some sense of a pulse on the market to be able to create good systems most of the time
29:44
unless you're going you know full Arbitrage route then sometimes you can I've seen some people do some crazy
29:49
arbitrages in like the Indian markets and things like that um unless you're going that route then maybe it can work
29:55
but most of the time you do want some kind of discretion experience sure no that's great um are you familiar with uh Christen kagi by any chance yeah I am I
30:02
I I've liked a lot of his content never looked into it too much because I'm not so much of a swing Trader but a lot of his content seems very good yeah and and
30:09
the reason why I'm bringing him up is he I do watch his stream occasionally and he mentioned that uh multiple people have contacted him to uh create his
30:17
system in an algo and he just says no you just can't do it right uh just because there's you know the market Cycles in Psychology and maybe um and
30:25
we'll bring that in the next question Neal Network eventually be a to kind of get there but at this point um you know
30:30
he says you know there's there's just a lot um that humans can do that algorithms cannot yet I'll I'll say with
30:38
that um it is it's a bit of a I've heard a
30:43
from a lot of traders that oh no my system can't be Quantified it can't be turned into an algo there's no way um
30:48
and what I've normally found is your main part of your strategy normally can be um a lot of the time I've I've rarely
30:55
found a circumstance where you can't turn the basic uh edge of what you're trying to exploit into an algo and
31:01
sometimes that can be really useful like I some of my most like common kind of I guess clients or or traders that come to
31:08
me for turning stuff into alos are you know trying to take a step back from the screens or having kids or things like that where in a time of your life it
31:15
really can make sense to maybe sacrifice some of those profits um to be able to run it in an algo and be able to take
31:20
time away and and kind of relax a bit at least from Trading um but obviously if you're in your kind of like Peak moment
31:26
of where you want to make the most am of gains sometimes yeah like The Human Experience can can beat it I agree I think that um most things can be
31:33
Quantified and then what that brings us uh to the next question is where do you see algorithmic trading and AI going um
31:40
you know do you feel like you know some of those discretionary parts are going to be able to be eliminated with neural
31:45
networks and things like that I think this is such an interesting question um I gave it some time to kind of think about
31:51
and my kind of conclusion at least from my stance obviously I don't know that much to be honest um is I don't think
31:58
much will really change I think a lot of the things that have already been um you know available to hedge funds and things
32:04
like that have been available for ages it's just kind of coming to the public uh markets I guess such as like Ai and being able to use it on Lower
32:11
performance uh computers and things like that um I can't really imagine a world where there's like no inefficiencies or
32:17
no Edge um I know that's some argument people have made based on you know alos get more into it but there's also
32:23
already hundreds of alos being run every single day and there's a lot of inefficiency you can take advantage of
32:28
um so realistic I don't see much changing I think maybe people if anything I I even saw a service recently
32:34
that was um using AI to create strategies and to be honest when I look at that on First Impressions obviously I
32:40
don't know much about the company is I'm just thinking overfitting there there's no way you're going to be able to just
32:45
crank out strategies with AI every single day for every single user that are going to be working it's just no way
32:52
um I've tried machine learning for a decent amount I'm I'm definitely no you know huge expert or anything like that but it is incredibly easy to overfit and
32:59
you know if you don't want to take my word you can look uh listen to people like Ernest Chan he's incredibly smart with machine learning and things like
33:04
that and even he says the main issues is you're probably going to overfit he doesn't even use machine learning for um
33:10
data mining or things like that he uses it for uh risk aversion and portfolio optimization which I think is a way
33:16
better use case um compared to trying to data mine through historical data I just don't think it's really a great route
33:22
for it at least yeah it's funny you mentioned that I did try to I had neural network trying to create these
33:27
strategies I had like reinforcement learning it did nothing but overfit I mean I still wanted to try it you know
33:32
what I mean but um but yeah I I totally agree with you the only use case at least at this point and again I'm not an
33:38
expert in in neural networks and machine learning but is was to actually have a strategy overlay and and essentially act
33:44
as a regime you know um uh filter you know essentially uh couldn't do it with
33:50
uh the actual strategy so totally agree and there's always going to be inefficiencies because if you think
33:56
about um you know earnings like um announcements things like that I mean
34:01
there's they're always going to exist and whoever even at that point executes more quickly is going to still be able
34:07
to capture it right there's just I guess I feel like the entire world that we're in I mean there's always going to be
34:13
these types of inefficiencies because you know we don't understand what the consumer is doing uh immediately I mean
34:19
maybe with blockchain and stuff like that that might be another situation but you said um you thinking about getting
34:24
into crypto too right potentially maybe potentially I I've considered it
34:29
obviously there's big advantages to crypto it's it's a lot easier on a data perspective running alos on it is really
34:35
easy um at least on like a coding standpoint and things like that a lot of accessible tools um main thing that
34:40
turns me away from crypto at the moment uh as we mentioned kind of commission commissions um I'm not a fan of paying a
34:46
huge amount of commissions um it can really hurt your EV um basically every trade you entering is an automatic loss
34:52
if your commissions are super high um so that you know makes it more difficult at least in my opinion opinion um there are
34:57
some really good kind of Arbitrage um things I've seen people kind of exploit um I just think if I was to move to
35:04
crypto it's a lot of effort it's a lot of energy a lot of new stuff I'd have to learn where if I probably applied that
35:10
time to my current systems or Futures markets where I'm bit better kind of have experience um I could probably do a
35:16
better result than trying to move all the way to crypto at the moment that could be my last words and maybe crypto is going to take us all over but uh Le
35:23
are the moment that's kind of my point of view on it yeah no I agree with you on the uh the the Commission in the
35:28
slipage it's it's pretty pretty challenging to say the least um but with that being said uh speaking of markets
35:34
uh where would you suggest a new uh Trader uh both discretionary algorithmic Trader start and and why that's a tough
35:41
question okay um for Edge I will say this again I'm probably going to you know people will come at me for saying
35:47
that I'm I'm diluting the edge in small caps but at least on a data perspective uh small caps is probably the easiest
35:52
way if you're going to go now that comes with risks that's why there's that inefficiency a lot of risks with small
35:57
caps lots of volatility risks um if you're just going for you know you want to go and deploy systems and you want to
36:03
try and develop systems I'll say probably Futures is the better way um key thing and I had a couple people ask me this recently is in Futures you're
36:10
not going to find um I had someone doing like a huge data analysis on NQ data and
36:15
I was like that at least in my opinion that's not how you're going to find an inefficiency in NQ like if you just look
36:21
at data mining on NQ um most of the time it's quite it's quite effective and if
36:26
got way bigger alos and way bigger institutions going over that data finding those inefficiencies of those Split Second decisions of what you can
36:33
kind of take advantage of um at least from my experience the best thing has been taking ideas like we mentioned from
36:39
books other Traders things like that and actually testing them um most of the time you can actually find some stuff that kind of sort of works and if you
36:44
find enough of those you can build a pretty good portfolio um doesn't need to be a crazy Edge that you're exploiting
36:49
um so yeah if you're going Algos probably Futures if you're going for a really clear Edge then probably uh small
36:54
caps um I've never ventured into Forex or any of those kind of things so I can't really speak on that um I think
37:00
also for for swing trading um large Cap's going to be a really good kind of tool there's a lot of really simple
37:06
strategies um I think one that comes off the top of my head is like a Christmas strategy um there's a lot of weird
37:11
inefficiencies like that we're in Christmas I think on average like it doubles the performance or whatever if
37:16
you're swing trading most stocks um just because there's normally a Christmas kind of rally obviously not huge sample
37:21
size so could be kind of overfit but at least for the past like I think 14 15 years it's worked quite well um so if
37:27
you're going for simple edges like that sometimes large caps and daily balls can be pretty good how did you get started you start doing algorithmic trading um
37:34
you know in the evening like how did you obviously the markets are only open certain amount of time so then you know
37:40
what's your thoughts there in regards to um you know lifestyle and working and having a job and also trading and maybe
37:46
you don't have that challenge because you don't have a family or anything like that yet um so you still have a pretty
37:51
long Runway but you know what's your thought process in regards to like timing and trading yeah I think this is
37:56
a really important question and quite a big question at least on it um I will say I have an unfair Advantage I don't
38:02
have any responsibilities um I I have to pay rent but apart from that uh not not
38:07
got a family or any know huge issues I haven't got a 9 to-5 um I will say when I started trading I did have parttime
38:13
jobs uh for a small bit I had a 9 to-5 job uh to kind of help build up some Capital um I can only speak from other
38:20
people's experiences here that I've kind of talked to as Traders some Traders I will say you really don't need to leave
38:25
your job um most of the time for for trading um there can be a lot of actually good edges that can happen in
38:31
the afternoons or sometimes if you even have like a remote job or a bit more of a flexible job you can even take advantage of those kind of open
38:37
strategies um and things like that um so I'd say if you're starting out really don't try and risk it all for trading um
38:45
a lot of the time is actually just going to make your trading worse I will say you'll need to spend a lot of the time
38:50
um with education of some some sort whether that's courses books you know all these kind of different me methods
38:56
you can go down um so I'd say it just takes a lot of time I I like it it
39:02
sounds bad but I think probably the first two two and a half years even nowadays most of the time I'm spending
39:07
at least eight hours seven hours a day on some sort of uh trading kind of topic
39:13
uh whether that's be I'm researching new ideas doing some data analysis improving my own systems trading um all those kind
39:20
of things really do take a lot of time um and if you can't devote that time then I'd say try and start small you know try and start maybe with some
39:26
string strategies if that works better if you um dip your toes into some very simple strategies that you can kind of get started with um but at the end of
39:32
the day at some point you're going to have to dedicate a decent amount of time whether that's staying up later or things like that um at least that's what
39:38
I've kind of gone about no that's that's that's great advice and and I agree you don't want to just jump uh uh full swing
39:45
into into it right um at least that's my experience unfortunately I did the opposite I ended up selling my business
39:51
and said oh I can do this and then you know it's a lot more challenging and then you have responsibilities too so so
39:56
anyone that thinks that um you you know obviously the Market's a great place to be I mean and I've was investing in
40:02
trading forever but um beforehand um and I maybe someday I'll talk about my story but um it is one of
40:10
those things where you want to try to keep your income and and do things on the side before and before jumping full
40:16
into it so just have to to warn the audience on that um so in regards to uh
40:21
learning resources uh I know you mentioned some books and and things like that what would you be um you know if
40:27
someone was starting from scratch uh and wanted to become a an algorithmic Trader
40:32
um obviously we talked about you know learning how to trade a little bit but you know what were the key resources or
40:37
mentors that you had that you would recommend um someone uh potentially look up or follow sure um I think I I've put
40:45
through a couple lists on my Twitter of the kind of best Traders I follow um a lot of the time
40:51
as on trading if you sell a CO or anything like that people instantly jump at it um instantly think you're not
40:57
profitable and there's a lot of really good traders that I know sell some good content um and most of the time even if
41:03
it is I have a very weird perspective on this I I know some people won't agree with me on it but even if it is fake
41:08
info but it pushes you in the right direction uh or like a fake Guru but it's you know still decent information
41:13
and pushes you in the right direction it can still be really valuable I've learned some weird things from you know YouTube videos that would probably a
41:19
scam YouTube video um but the actual strategy was kind of a cool idea and I tested it and some things I kind of took
41:25
from it um at least for direct ones I'd probably you know advocate for um Pon I
41:31
love his content and I know him quite well now um he's got an amazing python course for people trying to get into
41:36
kind of algorithmic trading uh mainly he focuses on options I may be wrong with that but feel free to look through his
41:42
content he's got some amazing news articles as well um I think books are great ones I already mentioned uh kind
41:48
of the main ones so onus Chan are really good uh evidence-based technical analysis Robert car's books um Perry J
41:55
Kaufman he's got some great kind of trading system books um those are really good introductions um if I think if you went through all of those books you'd
42:01
probably know as much as I do at least at the moment um just because a lot of those have a lot of good core information um and then Twitter I'd say
42:08
Twitter is an amazing resource if you're interested in back testing you could check out my back testing boot camp um but at the moment I'm not running it
42:14
currently for this month um but apart from that probably go through go through Twitter I think there's a lot of amazing resources um for some people that I can
42:21
name drop uh the short bear pant um who else is was a really good that I like um
42:28
probably you just have to explore around but there are some really good Traders you you can find oh uh the one lance B SMB Capital those are some really good
42:34
ones you can kind of look into yeah fin twit is a wealth of information uh I you know I wasn't very active on social
42:40
media uh for quite some time and then uh you know I joined Twitter and uh it's crazy it literally has its own little
42:47
ecosystem of Traders and and sharing a lot of really valuable information um so that's great um okay so um one last
42:55
question here um um you know is there any unique Insight or Alpha drop you know maybe something critical to your
43:00
success or something that uh you know you find that's very important in trading that might be less well known um
43:07
I will say obviously because I've stealing most of my ideas maybe they're not you know super um like secret um but
43:14
one I'd say is at least for a general advice um you really don't need something that predicts the next trade
43:19
or something like that you don't need predictive power over singular of Trades you really just need predictive power over a large enough sample of Trades um
43:25
and that's a key thing I try and always remember uh especially when I'm going into I made the mistake last month with
43:30
um thinking I was very close to a 10 figure 10 figure uh five figure uh month
43:36
and I was placing a lot of conviction that hey I've got a high win rate these last two trades of the month they'll
43:41
probably be profitable I'll hit $10,000 for the month it'll be amazing uh and what happened was I took two um deceny
43:48
sized losses and I didn't get very close to it um and that's just kind of how that happens I know over a large enough amount of Trades I'll I'll be profitable
43:55
but you really can't put that predictive power onto that Singler trade um it will most of the time buy you in the ass and
44:01
it's not going to be a fun experience um another one I'll say is that people really I I I think most Traders will go
44:07
against this Theory but some people think there is an in like a fully efficient market I think it's like the eim I whatever Theory um at least from
44:16
what I've seen there's a lot of markets that have very clear inefficiencies um obviously clearest one is SM caps um
44:21
there's other ones like earnings you know plays can actually have really big inefficiencies that are quite easy to exploit um and you don't have to go into
44:28
trading thinking no the whole Market's efficient there's no way I'm can find anything um edges are actually probably
44:33
the not they're not the easiest thing to come around um but there's a lot of edges it's actually more harder to find
44:39
an edge that you can execute well um and even if you're you know even if you're using an algo to execute them if it's a
44:44
strategy that goes into you know really big draw Downs um most people aren't going to survive that i' I've rarely seen people will be able to fully handle
44:51
long draw Downs um I remember there was an interview someone that handled a draw down for like you know a couple months
44:58
um most people I know at least for myself I could not handle being a draw down for a couple months I would probably go crazy um I think I'm doing
45:04
everything wrong so there are a lot of edges that you can you know exploit and use it just really needs to fit your personality or you need to have very
45:10
high kind of mental resilience to be able to handle those draw Downs or those issues with those strategies um and what
45:16
actually is harder is finding strategies that fit your personality and work well where you can execute them consistently
45:21
execute them um because there is no 97% 99 % win rate strategy that you're never
45:27
going to have to take a loss on like losing is just a very big part of of trading um and sometimes even when you
45:32
find those High win rate strategies um they're very low sample um so if you're thinking about it in like a systematic
45:38
way or algo way um am I going to have that much conviction on a 97% win rate
45:43
strategy that's only had 18 samples probably not I it's not you could get
45:49
really lucky with that there's only 18 samples to for me to look at and analyze um so I think those are probably the
45:55
ones I'd probably recommend the most oh and for like an alpha drop or a variable that you should kind of pay attention to
46:01
um is time I I think time is such a underrated variable that people don't really look at um maybe it's because I'm
46:06
from small caps and time really plays a huge factor into Data analysis um but in other you know markets time can be a
46:13
huge one and the changes between you know uh 8:30 and 9:30 and 11 there's
46:18
going to be really big sentiment changes based off either you know classical news releases or things like that that you can kind of dive into um and there good
46:26
things you can check out on your strategy so that's that's fantastic um perfect so um I think this was a
46:32
fantastic conversation how should people get a hold of you yeah sure you can uh look me up on Twitter it's Ghost Talk
46:37
trades um it has a ghost talk as the bird it was a bird I was a fan of when I was younger so I thought not use it as
46:42
the icon um if you're interested in coding Services you can look up unbiased trading um that's where I do coding
46:48
Services specifically for Traders I'm just kind of like a software person basically just for Traders um and then
46:53
if you're interested in back testing maybe check out my back practicing boot camp um otherwise I post pretty much
46:59
almost I think four threads a week at the moment um on different topics that I'm interested what I'm learning with
47:04
back testing and things like that um that you can get for free just on my Twitter I appreciate the conversation
47:09
today I think it went excellent and uh I can't wait to uh continue uh chatting with you and uh we'll stay in touch okay
47:16
yeah sure thank you so much for having me on man this a huge
47:24
pleasure
#Commodities & Futures Trading
#Brokerages & Day Trading