Depends on the outcome. If the automation can ensure that we only sell when there is a profit even if small while we hold when not, then I may pay for it, sure.
Quantopian is looking for algo's that compliment their fund. You may have a strategy that outperforms the market but is highly correlated to an algo quantopian is already backing.
Your algo is also tied to the quantopian API and data that they provide. I'm not sure what the effort would be like to port a Quantopian algo to a platform like Interactive Brokers.
I would pay for the learning process. The price would depend on topic depths. If you share the lessons titles I would have no problem in giving numbers.
Are you asking because you're doing market research? If so I would strongly urge you to reconsider. If you're not doing market research you can probably disregard most of this comment.
Here are the realities of trading algorithms and trading automation:
1. Approximately no one trades competently - let alone profitably - outside of a trading firm. With very few exceptions the process of designing, implementing, testing, deploying, monitoring and improving trading algorithms is not a feasible job for a single person. This is aside from the problem of data acquisition, engineering and processing. There are people who profitably trade on their own, but they're typically veterans of a trading firm looking to grab a larger piece of the PnL on their own book. I don't know why there is this persistent belief that a lone individual is going to be automating trading competently - in the past I've compared this to thinking you could singlehandedly develop and/or maintain half of DuckDuckGo.
2. The firms which automate most of their trading - and this is increasingly the default - know how to do it and know how to acquire the talent requisite for maintaining that core competency. They like to keep things in-house where possible. They are extraordinarily secretive and don't like depending on third party infrastructure or tooling unless it's absolutely necessary. You are probably not going to develop an ultra-sticky product like Bloomberg, Nanex or kdb. A third party dependency that inserts itself into the risk management, analysis or execution pipeline is a liability. Moreover the intense secrecy means you will have difficulty talking to, or hearing from, employees of actual firms.
3. This is an incredibly crowded space to try and enter. Many companies have popped up which market products to firms looking to automate some or all of their process. They're typically not very successful in this effort for two reasons. The first is what I already mentioned about secrecy among firms. But just as importantly it's very difficult to develop a reliable trading process by hand, let alone to automate it. The customers you want to work with are savvy and most of the existing products are crap, so you'd need to have a very strong product and a very compelling case for why yours is different and useful.
To put this all together for you: by asking this question on Hacker News you're strongly selecting for answers from people who almost certainly do not trade professionally. If they believe they are competently and effectively trading on their own without professional experience, it is most likely because they have vast unknown unknowns and a limited definition of what "automated trading" means (for example, they might think they're doing automated trading because they use the Interactive Brokers API instead of manually placing orders). You can target that kind of crowd, but if you do so you're entering the B2C cottage industry of (de facto) day trading education, not the B2B industry of trading itself. Unless you're attempting to build a fintech hypergrowth startup like Robinhood you really don't want to do this in a B2C capacity; then you're optimizing for customers who have both limited money and limited experience they could use to get any more money.
Anyone can make money in the market. I have 0 experience working at any financial firm, and I can beat any one of those systems. If you dedicate yourself to something you can do it.
Of course I wasn't talking about you. I'm sure you can singlehandedly beat RenTec, DE Shaw, Two Sigma, Citadel, Bridgewater, PDT, Tower, TGS, Headlands, AQR and DWR; all without any professional experience in finance. And naturally it makes more sense for you to work with a vastly smaller amount of capital, on your own, instead of simply joining one of those firms. After all you routinely beat their archaic and inefficient systems. You're a walking argument for active trading instead of index investing! Buffett should check in with you.
But you have to admit your skill is extraordinary, and I'm not talking about people like you in my earlier comment. I'm talking about most people, who cannot do the things that you can do. They probably come easily to you but you're just gifted with exceptional talent and work ethic. Most people can't trade profitably even with all the resources of a firm behind them; that you can do it while making all those other firms look silly is astounding and remarkable. So please don't feel as though I was downplaying your impressive accomplishments.
Of course they should, tell them to contact me, I will teach them a thing or three. My info is in my profile. The issue is they wouldn't give me their capital because they are scared.
I have come across people who are almost as good as me as well. They don't work in finance, have no finance experience, but they beat the most of those brokerages and are self sufficient, and work from home without a boss.
Anyways, there are many people who are doing what I am doing, to think they are not is not well thought out. Everyone who claims only experts can win are just not doing themselves a favour and the person listening to them.
I.e. Ignore the noise and get to work. There is money to be made. Just open your eyes, there is money to be had.
Take a look at my comment history. This is not a throwaway account :)
And for what it's worth, I don't doubt you at all. Like I said, I fully believe you are as incredible as you say. But not all of us can so easily obviate the work of entire firms staffed with hundreds of people employed to improve their trading results. You should count yourself lucky!
I have started looking down this path where I need something that can do a few things.
1) I have an account with Questrade, I need to enter the same trades on Interactive Brokers once I enter on Questrade. Also, I have customers who want to follow my trades, so if they can hook in and buy and sell at the same time I do, based on the same account balances, it can work for me.
2) I have a system where I enter if the stock price is > X. I use conditional orders for this now, so it doesn't really need to be automated, however it would be nice and would save me time.
3) The exiting is not automated because it's based on market sentiment, however, I wouldn't want to insert my data into any system that can leak the data. I could go for something where I put in supports, and instead of placing a stop order, it monitors the stock to fall under a price, and then stop me out using a market order. The reason I need this is because I would rather not have my stops advertised to brokerages.
4) Monitor my profile and alert me when a stock moves 5%,10% using beeps. This would be read-only. Something I am looking to do already.
The biggest drawback, what happens if the system fails? Are you going to reimburse me? If I write my own code, it's my own fault. Trusting someone else's code is hard to do. Also, if something breaks, how fast will it be fixed?
In the end, what I am doing now is conditional orders with a trailing stop that I adjust accordingly.
For those of you reading this and thinking this person has any idea what they're talking about, I'd like to direct your attention to their website: http://marketgodfathers.com (as listed in their profile).
In particular, note that this titan of industry advertises a monthly subscription service for options trading 1) with absolutely no technical or nuanced financial terminology, 2) using words like "play" to refer to trading, and 3) with wealth-centric hype videos from The Wolf of Wall Street and American Gangster.
Not much. I think there are open source kits out there. Since I would be risking my own cash anyway, and doing all the config work. Unless you can provide some proven edge or advantage
Yes if it included options. Most trading platforms are only stock, futures, or forex. Stocks and the like you have to be 100% directional, options on the other hand give you way more strategies that would be great to be able to automate from a trading platform. Especially if you had the ability to back test strategies and update your models over time, and adjust the option strategies. Or even the ability to combine options with another asset class like stock, futures, or pairs trading.
If someone created a platform like this with its own scripting language and REST api it would be like nothing available today.
Reasons why you would create a new scripting language that nobody is familiar with rather than using an existing language that at least some people would be familiar with?
I guess I went that direction because Thinkorswim and most other platforms all developed their own scripting language, but reflecting now I think something simple that everyone knows like Python or JavaScript lib/sdk would be a better approach.
Yes. I personally feel like the trading space is ripe for someone to come along and kick over everyone's cans. Etrade, Schwab, TD, they all work but -- they're all just so clunky, and early 2000's, and old, and just...just so not desirable. As a 25-year old professional who doesn't have the time to research stocks/options, and doesn't want to park his money in an index fund, I desperately need a MODERN alternative. But sadly there is none (no, Robinhood is not an alternative, it's a casino in your pocket).
I takes me about 40 hours to do proper DD on a company.
It depends if I know the sector well, if not then even more.
The actual trade is just hitting the sell/buy button.
That's once in a month.
I don't think I could ever automate the DD process.
You can come up with good AI and amazing algos but at the end of the day, a doctor knows more about the ill person just by looking at him/her for a few seconds, I feel like it's the same with trading/investing.
If I automate the valuation, I lose the connection to a company and its numbers.
Another example would be insider transactions, there is a difference how you evaluate insider trading, it depends on the company and its market cap, the sector, and the management.
I don't even know how I could ever come up with good indicators for those issues.
Now there are also people who use TA, that's your clientele.
20 comments
[ 3.4 ms ] story [ 64.0 ms ] threadI believe that if your algorithms perform well enough, you can actually lease them out to investors.
Your algo is also tied to the quantopian API and data that they provide. I'm not sure what the effort would be like to port a Quantopian algo to a platform like Interactive Brokers.
Here are the realities of trading algorithms and trading automation:
1. Approximately no one trades competently - let alone profitably - outside of a trading firm. With very few exceptions the process of designing, implementing, testing, deploying, monitoring and improving trading algorithms is not a feasible job for a single person. This is aside from the problem of data acquisition, engineering and processing. There are people who profitably trade on their own, but they're typically veterans of a trading firm looking to grab a larger piece of the PnL on their own book. I don't know why there is this persistent belief that a lone individual is going to be automating trading competently - in the past I've compared this to thinking you could singlehandedly develop and/or maintain half of DuckDuckGo.
2. The firms which automate most of their trading - and this is increasingly the default - know how to do it and know how to acquire the talent requisite for maintaining that core competency. They like to keep things in-house where possible. They are extraordinarily secretive and don't like depending on third party infrastructure or tooling unless it's absolutely necessary. You are probably not going to develop an ultra-sticky product like Bloomberg, Nanex or kdb. A third party dependency that inserts itself into the risk management, analysis or execution pipeline is a liability. Moreover the intense secrecy means you will have difficulty talking to, or hearing from, employees of actual firms.
3. This is an incredibly crowded space to try and enter. Many companies have popped up which market products to firms looking to automate some or all of their process. They're typically not very successful in this effort for two reasons. The first is what I already mentioned about secrecy among firms. But just as importantly it's very difficult to develop a reliable trading process by hand, let alone to automate it. The customers you want to work with are savvy and most of the existing products are crap, so you'd need to have a very strong product and a very compelling case for why yours is different and useful.
To put this all together for you: by asking this question on Hacker News you're strongly selecting for answers from people who almost certainly do not trade professionally. If they believe they are competently and effectively trading on their own without professional experience, it is most likely because they have vast unknown unknowns and a limited definition of what "automated trading" means (for example, they might think they're doing automated trading because they use the Interactive Brokers API instead of manually placing orders). You can target that kind of crowd, but if you do so you're entering the B2C cottage industry of (de facto) day trading education, not the B2B industry of trading itself. Unless you're attempting to build a fintech hypergrowth startup like Robinhood you really don't want to do this in a B2C capacity; then you're optimizing for customers who have both limited money and limited experience they could use to get any more money.
But you have to admit your skill is extraordinary, and I'm not talking about people like you in my earlier comment. I'm talking about most people, who cannot do the things that you can do. They probably come easily to you but you're just gifted with exceptional talent and work ethic. Most people can't trade profitably even with all the resources of a firm behind them; that you can do it while making all those other firms look silly is astounding and remarkable. So please don't feel as though I was downplaying your impressive accomplishments.
I have come across people who are almost as good as me as well. They don't work in finance, have no finance experience, but they beat the most of those brokerages and are self sufficient, and work from home without a boss.
Anyways, there are many people who are doing what I am doing, to think they are not is not well thought out. Everyone who claims only experts can win are just not doing themselves a favour and the person listening to them.
I.e. Ignore the noise and get to work. There is money to be made. Just open your eyes, there is money to be had.
Edit: deleted references to a throwaway account.
And for what it's worth, I don't doubt you at all. Like I said, I fully believe you are as incredible as you say. But not all of us can so easily obviate the work of entire firms staffed with hundreds of people employed to improve their trading results. You should count yourself lucky!
1) I have an account with Questrade, I need to enter the same trades on Interactive Brokers once I enter on Questrade. Also, I have customers who want to follow my trades, so if they can hook in and buy and sell at the same time I do, based on the same account balances, it can work for me.
2) I have a system where I enter if the stock price is > X. I use conditional orders for this now, so it doesn't really need to be automated, however it would be nice and would save me time.
3) The exiting is not automated because it's based on market sentiment, however, I wouldn't want to insert my data into any system that can leak the data. I could go for something where I put in supports, and instead of placing a stop order, it monitors the stock to fall under a price, and then stop me out using a market order. The reason I need this is because I would rather not have my stops advertised to brokerages.
4) Monitor my profile and alert me when a stock moves 5%,10% using beeps. This would be read-only. Something I am looking to do already.
The biggest drawback, what happens if the system fails? Are you going to reimburse me? If I write my own code, it's my own fault. Trusting someone else's code is hard to do. Also, if something breaks, how fast will it be fixed?
In the end, what I am doing now is conditional orders with a trailing stop that I adjust accordingly.
In particular, note that this titan of industry advertises a monthly subscription service for options trading 1) with absolutely no technical or nuanced financial terminology, 2) using words like "play" to refer to trading, and 3) with wealth-centric hype videos from The Wolf of Wall Street and American Gangster.
Run, don't walk, from this person's advice.
[1] https://www.reddit.com/user/marketgodfather/comments/8vm9jt/...
If someone created a platform like this with its own scripting language and REST api it would be like nothing available today.
I takes me about 40 hours to do proper DD on a company. It depends if I know the sector well, if not then even more. The actual trade is just hitting the sell/buy button.
That's once in a month.
I don't think I could ever automate the DD process.
You can come up with good AI and amazing algos but at the end of the day, a doctor knows more about the ill person just by looking at him/her for a few seconds, I feel like it's the same with trading/investing.
If I automate the valuation, I lose the connection to a company and its numbers. Another example would be insider transactions, there is a difference how you evaluate insider trading, it depends on the company and its market cap, the sector, and the management.
I don't even know how I could ever come up with good indicators for those issues.
Now there are also people who use TA, that's your clientele.