The Entrepreneur Forum | Financial Freedom | Starting a Business | Motivation | Money | Success

Welcome to the only entrepreneur forum dedicated to building life-changing wealth.

Build a Fastlane business. Earn real financial freedom. Join free.

Join over 80,000 entrepreneurs who have rejected the paradigm of mediocrity and said "NO!" to underpaid jobs, ascetic frugality, and suffocating savings rituals— learn how to build a Fastlane business that pays both freedom and lifestyle affluence.

Free registration at the forum removes this block.

Stock Trading software attempt

A detailed account of a Fastlane process...

LittleWolfie

Silver Contributor
Read Fastlane!
Speedway Pass
User Power
Value/Post Ratio
56%
Jun 28, 2018
951
531
Holbeach Hurn
So I built some software to use some unique indicators to provide a stock prediciton model, and tested it out. Certainly tiem and scale, since that's just IT. A good barrier as nobody else has the forumula I do.

I figured there was a need given how many there are and how popular they are, so I knocked it up shared them on reddit and set up a basic website. (Got traction and interest, but no wanted to buy so obviosuly no need. My version 2 attempt was to set up as a CD/DVD (see images and flog on ebay,etc) Give up the C for more £. No takers on gumtree/craigslist etc and ebay won't even list it :-(

This is the portfolio of shares my algorithm choose on the first run through.
Portfolio
TickerPrice% chgB/S$Shares$3,500$102,525
Cash$22,831
APH$87.27 (-0.53)-0.60%$86.74100$53$8,727
ATTO$7.75 (-0.30)-3.73%$7.9100$-15$775
BDX$223.73 (-1.87)-0.83%$213.58100$1,015$22,373
CAL$36.05 (0.16)0.45%$32.36100$369$3,605
CRY$26.90 (-0.05)-0.19%$19.65100$725$2,690
DIAX$19.66 (-0.18)-0.91%$17.59100$207$1,966
ESV$6.38 (-0.34)-5.06%$4.83100$155$638
KDMN$3.60 (-0.06)-1.64%$4.49100$-89$360
LITB$2.10 (0.01)0.48%$2.35100$-25$210
LW$65.43 (0.23)0.35%$54.57100$1,086$6,543
MO$55.63 (-0.26)-0.47%$59.55100$-392$5,563
MOGA$84.00 (0.17)0.20%$81.09100$291$8,400
NCLH$54.13 (0.46)0.86%$52.29100$184$5,413
NCZ$5.98 (-0.02)-0.33%$5.96100$2$598
NVG$14.91 (-0.01)-0.07%$14.36100$55$1,491
NYT$23.35 (0.35)1.52%$23.3100$5$2,335
PGEM$21.65 (0.00)0.00%$21.6100$5$2,165
RBS$7.78 (-0.09)-1.14%$7.505100$28$778
RYAM$18.23 (-0.08)-0.44%$19.69100$-146$1,823
STWD$21.77 (-0.04)-0.18%$21.075100$70$2,177
TGNA$10.64 (0.05)0.47%$11.46100$-82$1,06
Comments
Date

I think the date is, significant there is separate results for day trading and for long term buy and hold. The latter is what I trained my system to wards, but that’s just a function of my personal choice.
Timing
Due to time zone I’m not executing orders at standard US hours, so unsure if this significant or not, best results were buy at open, sell just before close.
Sector
I was only able to include US stocks, so there is a heavy tilt towards issues within the US economy.
Risk
There is no leverage, and no stop loss included. As you can see I also balanced on quantity rather than share value (a real commercial algro would likely adjust it’s position based on the value of the share, but putting 100 in all orders was an easy hack) in fact a trader could do this based on the stocks picked above.



example

site
 
Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

Attachments

  • $_57 (1).jpg
    $_57 (1).jpg
    158.7 KB · Views: 6
  • $_57 (1).png
    $_57 (1).png
    132.6 KB · Views: 7
  • $_57.jpg
    $_57.jpg
    171.2 KB · Views: 6
  • $_57.png
    $_57.png
    220.8 KB · Views: 8
  • disc.jpg
    disc.jpg
    158.7 KB · Views: 7

astr0

Grinding
Read Fastlane!
Read Unscripted!
Speedway Pass
User Power
Value/Post Ratio
140%
Dec 3, 2017
356
500
36
Lviv, Ukraine
Is it profitable? Do you have proofs?

That's the first thing I think when I see this kind of product for sale. Like an indicator or trading bot.

The reason is simple - if it's profitable why not start an investment fund, get a loan or just kick a$$ with your own/borrowed money? Why sell the golden goose that would definitely lose it's profitability once everyone starts using it (cause that's how the market works)?

If it's not profitable and not helpful in any other way then the value proposition is completely missing.

The stock trading industry also has quite a large authority entry barrier. Customers have to be sure that you're reliable, fast, precise and won't close the business in a few months.

After working for 12 years as a software developer in that industry and being involved in quite a few great projects, I still think it's nearly impossible entry. Maybe that doesn't matter that much for a small "casual investors", but definitely a thing for large trading companies, successful traders and funds who are the winners there and have all the money.

P. S. Small insight - traders have quite a problem choosing stocks to trade that isn't solved very well. It's as important as finding good entry and exit points.
 

LittleWolfie

Silver Contributor
Read Fastlane!
Speedway Pass
User Power
Value/Post Ratio
56%
Jun 28, 2018
951
531
Holbeach Hurn
Hey @astr0 , thanks for the reply. Some good info there.

Is it profitable?



Yes

Do you have proofs?



Define proof.



Well anyone was welcome to go to the website check the dates of publication and how the stocks did (by publishing them in advance, I could show results) , or to contact the third party simulator bot.





That's the first thing I think when I see this kind of product for sale. Like an indicator or trading bot.



The reason is simple - if it's profitable why not start an investment fund,



Lack of money, it is a lot cheaper to build software than start an investment fund. Something like 125,000 just to get a license.



get a loan or just kick a$$ with your own/borrowed money?



I have too little and erratic income to be eligible for a bank loan and a payday loan interest would wipe out the gains. Again what money? If I was on $40,000 a year then sure makes sense , I make about 1/10th that.



Why sell the golden goose that would definitely lose it's profitability once everyone starts using it (cause that's how the market works)?



Because A) Ithe money now is worth a lot more to me than the money in the future. I'm sure you know all about discounted cash flow ad marginal utility. My lower income means dollars now are worth a lot more. I will gladly pay Tuesday for a hamburger now. , B) the money can be invested in the servers and further development resulting in the potential of even greater gains and C) I can use the gains to build another one before it becomes a problem (also no reason I must sell it to the whole market. one portfolio to 200 traders @ £30 a month =£32,000 a year for expenses and 40k a year pre-tax profi that is enough for me to move to somewhere better, work on another project, and fund my other life goals plus D) it can work on a different basket of stocks per customer (or customer subset) to reduce the butterfly effect.

Once I have found a way to get my basic income, then I will perhaps use Darwin or similar with some of my own money.



If it's not profitable and not helpful in any other way then the value proposition is completely missing.



It made a profit.

The stock trading industry also has quite a large authority entry barrier. Customers have to be sure that you're reliable, fast, precise and won't close the business in a few months.



I suppose that authority entry is a bit like marketing. It is certainly impossible to show that I would be open for a long time(too little runway, which is why I had to close it down).

The processing speed is far from fast but that is just a matter of throwing more GPU Rigs (and $,£,€) at it.

I saw no reason to worry about clock speed since I could spend all weekend running the predictions. (which meant I could use cheaper delayed data and fewer server credits) Essentially I optimized for predictions further ahead than for quick predictions



[/Quote]



After working for 12 years as a software developer in that industry and being involved in quite a few great projects, I still think it's nearly impossible entry.

Maybe that doesn't matter that much for a small "casual investors", but definitely a thing for large trading companies, successful traders and funds who are the winners there and have all the money.



This should perhaps be my 3rd or 4th business.



P. S. Small insight - traders have quite a problem choosing stocks to trade that isn't solved very well. It's as important as finding good entry and exit points.


I know, in what way does my system avoid doing that?



on Mar 21st prior to market open my Algo has sell at high on;

KNOPKNOT Offshore Partners LP
LWLamb Weston Holdings, Inc.
LITBLightInTheBox Holding Co., Ltd.
And others

MOSCMosaic Acquisition Corp.
NTZNatuzzi, S.p.A.
NEWRNew Relic, Inc.

These could have been purchased/optioned the week before (when they were indicated as buy low)

In fact with enough money, the entire process could be automated. So no input/decision whatsoever would be needed for the trader.
 

ChrisV

Legendary Contributor
EPIC CONTRIBUTOR
Read Fastlane!
Read Unscripted!
Speedway Pass
User Power
Value/Post Ratio
225%
May 10, 2015
3,141
7,055
Islands of Calleja
I don't know how to read these numbers.. did it work?
 
Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

ChrisV

Legendary Contributor
EPIC CONTRIBUTOR
Read Fastlane!
Read Unscripted!
Speedway Pass
User Power
Value/Post Ratio
225%
May 10, 2015
3,141
7,055
Islands of Calleja
Do you have the ability of Backtest stock hypotheses? Like if I gave you model would you be able to back and see if it works?
 
Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

broswoodwork

Intermediate User of the Flying Guillotine
Read Fastlane!
Read Unscripted!
Speedway Pass
User Power
Value/Post Ratio
313%
Oct 16, 2015
890
2,790
How can I prove it to you if your unable to read the numbers/stocks?
Walk him through it as simply as possible. Infograph, etc. Having something that does some technical thing is neat; selling it successfully to non-technical people is a business.

Think Wozniak sitting in the back of the Computer Club playing with his plywood computer vs Jobs swirling into a decision maker's office eating a piece of fruit and breaking down how this contraption would eliminate typing pools.
 

ChrisV

Legendary Contributor
EPIC CONTRIBUTOR
Read Fastlane!
Read Unscripted!
Speedway Pass
User Power
Value/Post Ratio
225%
May 10, 2015
3,141
7,055
Islands of Calleja
Walk him through it as simply as possible. Infograph, etc. Having something that does some technical thing is neat; selling it successfully to non-technical people is a business.

Think Wozniak sitting in the back of the Computer Club playing with his plywood computer vs Jobs swirling into a decision maker's office eating a piece of fruit and breaking down how this contraption would eliminate typing pools.
People don't realize how profound and important this concept is. Apple's entire business model was taking ultra-complicated things and making them so simple your grandma could use them. Steve Jobs said "You want to find an industry that's ripe for disruption? Find a product that's too complicated and simplify it."
 

astr0

Grinding
Read Fastlane!
Read Unscripted!
Speedway Pass
User Power
Value/Post Ratio
140%
Dec 3, 2017
356
500
36
Lviv, Ukraine
Because A) Ithe money now is worth a lot more to me than the money in the future. I'm sure you know all about discounted cash flow ad marginal utility. My lower income means dollars now are worth a lot more. I will gladly pay Tuesday for a hamburger now. , B) the money can be invested in the servers and further development resulting in the potential of even greater gains and C) I can use the gains to build another one before it becomes a problem (also no reason I must sell it to the whole market. one portfolio to 200 traders @ £30 a month =£32,000 a year for expenses and 40k a year pre-tax profi that is enough for me to move to somewhere better, work on another project, and fund my other life goals plus D) it can work on a different basket of stocks per customer (or customer subset) to reduce the butterfly effect.
Agree on everything except 200 traders @ £30. This software should be much more expensive if it really works. Think of value is can provide managing large sums of money.

I'm still a bit suspicious cause I have seen profitable traders, but haven't seen profitable automated trading software that wasn't using tech cheats on performance basically processing information and sending orders faster than other or leveraging human traders, mimicking some of their actions while slightly adjusting them.

However, I truly believe that if there are profitable traders their strategies can be automated too. Main issues with that are that they are not willing to share them or it turns out to be very subjective and hard to express even in human words, not formal code.
 
Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

Tourmaline

Gold Contributor
Read Fastlane!
Read Unscripted!
Summit Attendee
Speedway Pass
User Power
Value/Post Ratio
120%
Jun 4, 2019
898
1,082
Texas
People not buying is not necessarily because of lack of need. I would venture it's a lack of trust.

Your website is a .blog address? I'd run far away based on that alone.

Beyond that don't expect people to trust you because of 1 year of positive ROI. I would want to see at least 5 years. If this is truly legit, then you will want to some patience and figure out an income model that matches the expectations of your audience. Confused about the $30 for the software, that's super cheap. But your site is saying $12,000/yr and $2,000/lifetime which also hurts my head.

You will want to convince me I'm not better off simply buying the whole american market in an index fund.

I'm intrigued, and if this is actually profitable in the long run, you should be able to make a killing too!

Do you have more pictures of the actual software in use?
 

ChrisV

Legendary Contributor
EPIC CONTRIBUTOR
Read Fastlane!
Read Unscripted!
Speedway Pass
User Power
Value/Post Ratio
225%
May 10, 2015
3,141
7,055
Islands of Calleja
Beyond that don't expect people to trust you because of 1 year of positive ROI. I would want to see at least 5 years.
Bingo... can you backtest this method using a larger sample?

@LittleWolfie, do you know how to backtest hypotheses? Let's say I come up with an idea. "I think that if you picked stocked based off _____, it would be profitable" then going back in history to test that prediction. Do you know how to do that?
 

Kak

Legendary Contributor
FASTLANE INSIDER
EPIC CONTRIBUTOR
Read Rat-Race Escape!
Read Fastlane!
Read Unscripted!
Summit Attendee
Speedway Pass
User Power
Value/Post Ratio
493%
Jan 23, 2011
9,678
47,666
34
Texas
Bingo... can you backtest this method using a larger sample?

@LittleWolfie, do you know how to backtest hypotheses? Let's say I come up with an idea. "I think that if you picked stocked based off _____, it would be profitable" then going back in history to test that prediction. Do you know how to do that?

Many have tried to use math and algorithms and theories to predict the future... On average, over the years, very few beat a simple index fund.

There are three kinds of money in the stock market... Smart, Dumb and Lazy.

Dumb: The battle cry of dumb money is "OMG this is the next big thing" as if they are the only one with a really popular idea. They act irrespective of ANY understanding of the value they get for their money. They also like to say "location, location, location" a lot when they are buying personal homes that they call investments. Most dumb money people also have some lazy money. "I'm buying only pot stocks, it is going to be HUUUUUUGE."

Lazy: They only bitch when the DOW goes down. They hand their money over to a 26 year old kid with a big well known firm and pay 2% maintenance fees annually for the privilege of being diversified exactly like everyone else on earth. It is so "diversified" in fact that the upside is almost nothing. They would almost always be better served tossing it all in a macro market index fund and keeping their hands off of it because then they wouldn't realize their gains. This money moves slow and stays invested though volatility and helps the stability of the markets for the smart money. They make 8 percent a year. This is the consensus. "I don't want all my eggs in one basket." "I maxed out my 401k and IRA this year." :rofl:

Smart: Knows what makes a prudent investment. They know to make relative money in the markets you need to bet against the consensus and be right. They have wins and losses, but weed out the stupid crap that the lazy money is in. They are there to sell stock to the dumb money. They generally beat the market, but sometimes, less often than they win, the market beats them, and that is OK, they understand that it can happen. Over a lifetime, they make a LOT more money. "I think this is a good buy because it is inexpensive relative to competitors and I believe they have a stronger position because of X, Y, and Z. The price to earnings ratio indicates that others have not caught on to this yet, so I am going to invest an intelligent amount of my portfolio into it."

My point with all of this... Just be smart money. Know why you are doing the things you do and you will be ahead of the curve as an investor. No one has ever captured smart money in an AI that I know about... Because if they did, it would be the new lazy consensus. ;)

@LittleWolfie no offense, but if JP Morgan, Merrill Lynch, Goldman Sachs, BNY Melon, and any number of other funds with literally trillions under management haven't cracked this code, neither have you. As @ChrisV said, the data is available for a back test. That would be a good start.
 
Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.
Last edited:

ChrisV

Legendary Contributor
EPIC CONTRIBUTOR
Read Fastlane!
Read Unscripted!
Speedway Pass
User Power
Value/Post Ratio
225%
May 10, 2015
3,141
7,055
Islands of Calleja
My point with all of this... Just be smart money. Know why you are doing the things you do and you will be ahead of the curve as an investor. No one has ever captured smart money in an AI that I know about... Because if they did, it would be the new lazy consensus.
No, I agree. Here's my overall philosophy on data: data should almost never be used by itself to predict almost anything. It should be one tool in a toolkit. You still have to know what you're doing. But having data can be a great way of testing hypothesis.

@LittleWolfie no offense, but if JP Morgan, Merrill Lynch, Goldman Sachs, BNY Melon, and any number of other funds with literally trillions under management haven't cracked this code, neither have you.
I'm not sure I agree. I think there are always better ways of doing things, and regular people with a fresh perspective can sometimes come up with those ideas. I mean we can sit here all day coming up with ideas of how to do things better than big companies.

I'd be interested in seeing this tested on a much bigger sample. Picking 20 stocks definitely has the potential for sampling bias. Run that test 10 or 20 more times, at least.

I'd also be interested in knowing what criteria he's using to pick them, although I'm not sure he'd be willing to share that.
 

ChrisV

Legendary Contributor
EPIC CONTRIBUTOR
Read Fastlane!
Read Unscripted!
Speedway Pass
User Power
Value/Post Ratio
225%
May 10, 2015
3,141
7,055
Islands of Calleja
And I'm not saying people should just let some AI bot pick stocks for them. I think that AI at this point would be good for just flagging potential opportunities. In other words, letting AI augment your decision making, rather than replacing it. I don't think that most AI is at the point where it can make decisions for us, especially not market decisions.

Anyway I'd be curious as to the results of about 20 more backtests to see if it still holds.
 

astr0

Grinding
Read Fastlane!
Read Unscripted!
Speedway Pass
User Power
Value/Post Ratio
140%
Dec 3, 2017
356
500
36
Lviv, Ukraine
And I'm not saying people should just let some AI bot pick stocks for them
That is actually quite possible or even happening already. On of the main benefits of the last trading platform that I were coding on 9-5 was helping to filter out those stocks.
Basically it done ton of calculations based on market data, news and historical data and the trader configured sorting, filtering and range. It also allowed to quickly dig into that data and set custom formulas like in Excel.
Some traders used other, more familiar, software to send orders, but pretty much all of the 250+ people in the client's company used that software to find where to put their attention.
Quite possibly a simple AI could pick up their usage patterns and filter the list even further.
 
Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

ChrisV

Legendary Contributor
EPIC CONTRIBUTOR
Read Fastlane!
Read Unscripted!
Speedway Pass
User Power
Value/Post Ratio
225%
May 10, 2015
3,141
7,055
Islands of Calleja
That is actually quite possible or even happening already. On of the main benefits of the last trading platform that I were coding on 9-5 was helping to filter out those stocks.
Basically it done ton of calculations based on market data, news and historical data and the trader configured sorting, filtering and range. It also allowed to quickly dig into that data and set custom formulas like in Excel.
Some traders used other, more familiar, software to send orders, but pretty much all of the 250+ people in the client's company used that software to find where to put their attention.
Quite possibly a simple AI could pick up their usage patterns and filter the list even further.
Well I'll give a similar example from another industry. Prison reform. In New Jersey incorporating AI-derived risk scores has helped reduce violent crime by significant margins. Murder, assault. I linked to the part where she talks about the tool, but the talk is worth a watch. You obviously can't make a parole decision just based off AI, but a score like this gives judges a potential recidivism and potential for violent crime score that they can now incorporate into their decisions.

1568306163851.png

1568306189259.png

Anne Milgram: Why smart statistics are the key to fighting crime (10:25)
 

Kak

Legendary Contributor
FASTLANE INSIDER
EPIC CONTRIBUTOR
Read Rat-Race Escape!
Read Fastlane!
Read Unscripted!
Summit Attendee
Speedway Pass
User Power
Value/Post Ratio
493%
Jan 23, 2011
9,678
47,666
34
Texas
And I'm not saying people should just let some AI bot pick stocks for them. I think that AI at this point would be good for just flagging potential opportunities. In other words, letting AI augment your decision making, rather than replacing it. I don't think that most AI is at the point where it can make decisions for us, especially not market decisions.

Anyway I'd be curious as to the results of about 20 more backtests to see if it still holds.
Even so. If AI was that advanced, it would become the new lazy money and the effects would get watered back down to the neutered reflection of macro economic performance. Like competition, when everyone has the same idea it waters down the profitability.

Then you are back to smart money. Doing things differently than lazy money is profitable.

Stocks, real estate, entrepreneurship... Profitability, notwithstanding oligarchical enterprise, is a moving target.
 

Kak

Legendary Contributor
FASTLANE INSIDER
EPIC CONTRIBUTOR
Read Rat-Race Escape!
Read Fastlane!
Read Unscripted!
Summit Attendee
Speedway Pass
User Power
Value/Post Ratio
493%
Jan 23, 2011
9,678
47,666
34
Texas
Well I'll give a similar example from another industry. Prison reform. In New Jersey incorporating AI-derived risk scores has helped reduce violent crime by significant margins. Murder, assault. I linked to the part where she talks about the tool, but the talk is worth a watch. You obviously can't make a parole decision just based off AI, but a score like this gives judges a potential recidivism and potential for violent crime score that they can now incorporate into their decisions.

View attachment 27371

View attachment 27372

Anne Milgram: Why smart statistics are the key to fighting crime (10:25)

Look up the Arnold Foundation and their work on pre-trial risk assessment. This is EXACTLY what they are trying to do.
 
Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

Schonox

Contributor
Read Unscripted!
User Power
Value/Post Ratio
92%
May 25, 2019
72
66
From what I know and think about this subject, I don't think it can ever work.
It's an algorythm, it cannot take into account human decisions like FED rate cuts, board decisions, trade wars etc.
It may work in the short term but eventually it will fail, no doubt.
How does it measure a stock's value?
Sorry...
 

Kevin88660

Platinum Contributor
FASTLANE INSIDER
Read Unscripted!
Speedway Pass
User Power
Value/Post Ratio
118%
Feb 8, 2019
3,457
4,078
Singapore
So I built some software to use some unique indicators to provide a stock prediciton model, and tested it out. Certainly tiem and scale, since that's just IT. A good barrier as nobody else has the forumula I do.

I figured there was a need given how many there are and how popular they are, so I knocked it up shared them on reddit and set up a basic website. (Got traction and interest, but no wanted to buy so obviosuly no need. My version 2 attempt was to set up as a CD/DVD (see images and flog on ebay,etc) Give up the C for more £. No takers on gumtree/craigslist etc and ebay won't even list it :-(

This is the portfolio of shares my algorithm choose on the first run through.
Portfolio
TickerPrice% chgB/S$Shares$3,500$102,525
Cash$22,831
APH$87.27 (-0.53)-0.60%$86.74100$53$8,727
ATTO$7.75 (-0.30)-3.73%$7.9100$-15$775
BDX$223.73 (-1.87)-0.83%$213.58100$1,015$22,373
CAL$36.05 (0.16)0.45%$32.36100$369$3,605
CRY$26.90 (-0.05)-0.19%$19.65100$725$2,690
DIAX$19.66 (-0.18)-0.91%$17.59100$207$1,966
ESV$6.38 (-0.34)-5.06%$4.83100$155$638
KDMN$3.60 (-0.06)-1.64%$4.49100$-89$360
LITB$2.10 (0.01)0.48%$2.35100$-25$210
LW$65.43 (0.23)0.35%$54.57100$1,086$6,543
MO$55.63 (-0.26)-0.47%$59.55100$-392$5,563
MOGA$84.00 (0.17)0.20%$81.09100$291$8,400
NCLH$54.13 (0.46)0.86%$52.29100$184$5,413
NCZ$5.98 (-0.02)-0.33%$5.96100$2$598
NVG$14.91 (-0.01)-0.07%$14.36100$55$1,491
NYT$23.35 (0.35)1.52%$23.3100$5$2,335
PGEM$21.65 (0.00)0.00%$21.6100$5$2,165
RBS$7.78 (-0.09)-1.14%$7.505100$28$778
RYAM$18.23 (-0.08)-0.44%$19.69100$-146$1,823
STWD$21.77 (-0.04)-0.18%$21.075100$70$2,177
TGNA$10.64 (0.05)0.47%$11.46100$-82$1,06
Comments
Date

I think the date is, significant there is separate results for day trading and for long term buy and hold. The latter is what I trained my system to wards, but that’s just a function of my personal choice.
Timing
Due to time zone I’m not executing orders at standard US hours, so unsure if this significant or not, best results were buy at open, sell just before close.
Sector
I was only able to include US stocks, so there is a heavy tilt towards issues within the US economy.
Risk
There is no leverage, and no stop loss included. As you can see I also balanced on quantity rather than share value (a real commercial algro would likely adjust it’s position based on the value of the share, but putting 100 in all orders was an easy hack) in fact a trader could do this based on the stocks picked above.



example

site
I am in the space of financial sales.

Buyers need to know other keys factors in your performance and strategy.

Performance related
1) What is the performance of major US stock index in the same 1.25 year period?

2) Maximum monthly draw down within the 1.25year?

Strategy
1) Is this a buy only mean reversion strategy?

2) How many strategies are being employed at the same time in your portfolio?
 
Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

LittleWolfie

Silver Contributor
Read Fastlane!
Speedway Pass
User Power
Value/Post Ratio
56%
Jun 28, 2018
951
531
Holbeach Hurn
Do you have the ability of Backtest stock hypotheses? Like if I gave you model would you be able to back and see if it works?

It was backtested, all though I lack that ability. If I had the resources for the ML rig and enough back history in your model the I could pick things from the future of your model and let you validate it.


Walk him through it as simply as possible. Infograph, etc. Having something that does some technical thing is neat; selling it successfully to non-technical people is a business.

Think Wozniak sitting in the back of the Computer Club playing with his plywood computer vs Jobs swirling into a decision maker's office eating a piece of fruit and breaking down how this contraption would eliminate typing pools.

That is far from proving it. Okay, imagine this replacing a hedge fund manager.


People don't realize how profound and important this concept is. Apple's entire business model was taking ultra-complicated things and making them so simple your grandma could use them. Steve Jobs said "You want to find an industry that's ripe for disruption? Find a product that's too complicated and simplify it."

This requires access to a large number of grandmothers for user testing and identifying a product that is too complicated. Email?Internet?Facebook?

Agree on everything except 200 traders @ £30. This software should be much more expensive if it really works. Think of value is can provide managing large sums of money.
20 traders @ £300? Okay, seriously say 0.5% Assets under management + £300 pcm "management fees"
I'm still a bit suspicious cause I have seen profitable traders, but haven't seen profitable automated trading software that wasn't using tech cheats on performance basically processing information and sending orders faster than other or leveraging human traders, mimicking some of their actions while slightly adjusting them.

Sure, I can understand that. What about this even better one?

However, I truly believe that if there are profitable traders their strategies can be automated too. Main issues with that are that they are not willing to share them or it turns out to be very subjective and hard to express even in human words, not formal code.

If past trades are a matter of record, then you can (in theory) reverse engineer their trades, just like you can with software. Also AI can deal with fuzzy logic and other subjective things

People not buying is not necessarily because of lack of need. I would venture it's a lack of trust.

Your website is a .blog address? I'd run far away based on that alone.

Yes, lack of money hurts the operation. Initially I was hoping to attract a partner or too to gain legitimacy and funding/sales. End of the day it was a cool thing.

Beyond that don't expect people to trust you because of 1 year of positive ROI. I would want to see at least 5 years. If this is truly legit, then you will want to some patience and figure out an income model that matches the expectations of your audience.

Simulation was an average over decades,1 year was just what I did in advance to try and demo it. This should be left till I have at least 5 years runway then cool, good to know!

It might be decades away then, but in the meantime you can see it. Once, I have got my initial income going by hook or crook, be it websites/technical writing, a job or something else then I can look at self-funding this off the ground.


Confused about the $30 for the software, that's super cheap.

That was what most trading bots were going for, when I did the research. Some of them claim to be even better.


You will want to convince me I'm not better off simply buying the whole american market in an index fund.
I'm intrigued, and if this is actually profitable in the long run, you should be able to make a killing too!

Depends on if you want a better var and sharpe while self-managing or if you want to just set and forget in which case, buy the index fund.

Do you have more pictures of the actual software in use?

I might on my old phone, but the backend is text-only so pretty boring.

Bingo... can you backtest this method using a larger sample?

@LittleWolfie, do you know how to backtest hypotheses? Let's say I come up with an idea. "I think that if you picked stocked based off _____, it would be profitable" then going back in history to test that prediction. Do you know how to do that?

I have no idea how to backtest, but I had it backtested with a larger sample ( Upgrade Capital have it and the results) b

No, I agree. Here's my overall philosophy on data: data should almost never be used by itself to predict almost anything. It should be one tool in a toolkit. You still have to know what you're doing. But having data can be a great way of testing hypothesis.

I'm reminded of the story of the late mathematical student who thinking the whiteboard
work was homework solved an "impossible" theorem that had stumped professionals
for decades.

I'm not sure I agree. I think there are always better ways of doing things, and regular people with a fresh perspective can sometimes come up with those ideas. I mean we can sit here all day coming up with ideas of how to do things better than big companies.

I'd be interested in seeing this tested on a much bigger sample. Picking 20 stocks definitely has the potential for sampling bias. Run that test 10 or 20 more times, at least.

It has been, displaying the data in a better way needs to be accomplished.

I'd also be interested in knowing what criteria he's using to pick them, although I'm not sure he'd be willing to share that.

To be honest, I'm unsure myself.

And I'm not saying people should just let some AI bot pick stocks for them. I think that AI at this point would be good for just flagging potential opportunities. In other words, letting AI augment your decision making, rather than replacing it. I don't think that most AI is at the point where it can make decisions for us, especially not market decisions.

That's how mine works.


Anyway I'd be curious as to the results of about 20 more backtests to see if it still holds.
[/QUOTE]


From what I know and think about this subject, I don't think it can ever work.
It's an algorythm, it cannot take into account human decisions like FED rate cuts, board decisions, trade wars etc.
It may work in the short term but eventually it will fail, no doubt.
It failed during the recession during backtesting, and most algos are being killed by Brexit. However most losses are due to human error, so even basic algos can beat them because they can be programmed to avoid throwing good money after bad.


How does it measure a stock's value?
Sorry...

Actually it never bothers to measure value, it is sufficient to know that the trend is likely to be upwards until a week next Wednesday (Not an actual output)


I am in the space of financial sales.

Buyers need to know other keys factors in your performance and strategy.

Performance related
1) What is the performance of major US stock index in the same 1.25 year period?

2) Maximum monthly draw down within the 1.25year?

Strategy
1) Is this a buy only mean reversion strategy?

2) How many strategies are being employed at the same time in your portfolio?

@Kevin88660 Here are some numbers for you, hope they make more sense to you than to me!


Absolute Return: 4.75%

Sharpe Ratio: 4.49

Weekly Var 99: -1.86%

I only employed one strategy, It is a market maker "vampire" trader that aims to buy low and sell high, there is no mean reversion, no balancing, option's or anything else complicated.

That was an average rather than a specific year,I think the average over the same period was about 10%.
 

Kevin88660

Platinum Contributor
FASTLANE INSIDER
Read Unscripted!
Speedway Pass
User Power
Value/Post Ratio
118%
Feb 8, 2019
3,457
4,078
Singapore
@Kevin88660 Here are some numbers for you, hope they make more sense to you than to me!


Absolute Return: 4.75%

Sharpe Ratio: 4.49

Weekly Var 99: -1.86%

I only employed one strategy, It is a market maker "vampire" trader that aims to buy low and sell high, there is no mean reversion, no balancing, option's or anything else complicated.

That was an average rather than a specific year,I think the average over the same period was about 10%.
[/QUOTE]
The sharpe ratio is very good.

The draw down is small. You can make money using the strategy by employing a large leverage even if you do not have much capital.
 
Dislike ads? Remove them and support the forum: Subscribe to Fastlane Insiders.

LittleWolfie

Silver Contributor
Read Fastlane!
Speedway Pass
User Power
Value/Post Ratio
56%
Jun 28, 2018
951
531
Holbeach Hurn
@Kevin88660 Here are some numbers for you, hope they make more sense to you than to me!


Absolute Return: 4.75%

Sharpe Ratio: 4.49

Weekly Var 99: -1.86%

I only employed one strategy, It is a market maker "vampire" trader that aims to buy low and sell high, there is no mean reversion, no balancing, option's or anything else complicated.

That was an average rather than a specific year,I think the average over the same period was about 10%.

The sharpe ratio is very good.

The draw down is small. You can make money using the strategy by employing a large leverage even if you do not have much capital.
[/QUOTE]

Are you able to tell me the drawdown from those numbers I provided? I failed at calculating it. Can you give me an idea of drawdown numbers that are bad, ok,good and amazing?

For refining the algorithm.

Thanks, a good sharpe ratio is a common response,makes sense since I wad optimising for loss prevention over gains.

I know Berkshire hathaway has been <1

Would you be able to caluclate a "safe withdrawl rate" I'm still learning about leverage and options,but will definitely use this once I can afford a rig to run it and the stock brokerarge account
 

Kevin88660

Platinum Contributor
FASTLANE INSIDER
Read Unscripted!
Speedway Pass
User Power
Value/Post Ratio
118%
Feb 8, 2019
3,457
4,078
Singapore
You should have the drawdown figure if you have run it live?

A sharp ratio of 4-5 means most of the months you have positive return. And even you have down months the magnitude it will be a lot less than the up month.
 

Kevin88660

Platinum Contributor
FASTLANE INSIDER
Read Unscripted!
Speedway Pass
User Power
Value/Post Ratio
118%
Feb 8, 2019
3,457
4,078
Singapore
You should trade it Live on Interactive Broker.

You should learn the knowledge about trading and finance also along the way.

How old are you? Is your background in IT? If you are serious about this you should join a prop trading firm to learn the basic of the industry first.

What you are doing here is not selling a program. Because serious people in the business of trading do not buy a trading program to do the work for them. You are actually selling a trading strategy. The only people that are willing to pay for a trading strategy are senior management of big and established trading firms. They hired people with quantitative skills and pay them 200k to research on strategies. So your best bet is to work for them.
 

Post New Topic

Please SEARCH before posting.
Please select the BEST category.

Post new topic

Guest post submissions offered HERE.

Latest Posts

New Topics

Fastlane Insiders

View the forum AD FREE.
Private, unindexed content
Detailed process/execution threads
Ideas needing execution, more!

Join Fastlane Insiders.

Top