Back to Community
Questionnaire for Quantopian Live-Brokerage Traders

Hi y'all. I know a lot of us are scrambling due to the sudden discontinuation of Quantopian's live-trading brokerage integrations. I have some questions that I think would be useful in order to gauge the live-trading community, its needs, and where to go from here. Maybe this will help you identify others in the same boat as you -- companions for your new journey. And perhaps new solutions will be developed in response to what people write here.

If you'd like me to add more questions or anything like that just PM me and I'll update and add it to this post.

  1. What broker were you using?
  2. How long have you been live-trading?
  3. How long have you been working with the Quantopian platform?
  4. How much capital did you invest in your algo(s)?
  5. What kinds of returns did you see? Were they as strong as you expected?
  6. How deep is your finance/stock market background?
  7. How deep is your programming background? How comfortable are you with Python? What is your "native" or favorite language?
  8. Can you give a high-level description of your strategy/ies? Economic premise? Methods utilized? Amount and type of securities used?
  9. What Quantopian features were utilized to inform your trading signals? OHCL bars? Fundamental data? Alternative data? Daily or minutely? Pipeline?
  10. Were there any features missing from Quantopian that you wish you'd had access to?
  11. What do you think you'll do now in order to keep live-trading?
  12. What are you goals? What are you hoping to get out of algorithmic trading?
  13. What brought you to Quantopian?

There's another thread where people are thrashing out -- lets keep this one strictly positive+productive.

13 responses

I'll start.

  1. Brokerage: Robinhood
  2. Live-trading only a couple months.
  3. Been on Quantopian only a few months -- but I've been treating it like a full-time job.
  4. Hard to say, since I had buy-and-hold securities in my Robinhood account alongside the Q trading. Q was controlling maybe $25k
  5. Couldn't quantify how successful my Q trading was, due to the aforementioned issue.
  6. None. Never had an interest in it. I'm a "millennial" -- the stock market is evil.
  7. I actually have quite a lot of web development experience from some years back, specifically LAMP stack. But I only learned Python recently in order to use Quantopian.
  8. I've been finding most success exploring novel ways to exploit mean reversion characteristics. Using SMA and developing my own indicators, doing my own math. Also just coding up algos to automate exploiting some pretty obvious alpha. None of my more clever ideas so far have panned out.
  9. I've been mostly just working with minutely OHCL bars. Developed some algorithms with predefined basket of stocks, but they all improved upon using pipeline with fundamental screens.
  10. Only being able to pull data in via CSV is a bit limiting. Only being able to pull in data before market is a bit limiting. I haven't gotten there yet, but it seems machine learning needs more power than the cloud situation provides. I could also see being able to interact with or tune parameters on an algo in real-time -- there could be some potential there, you know some bio-feedback. I wish it had handled deposits/withdrawals elegantly. I wish it could handle multiple algos in one brokerage account and could compartmentalize what each algo was doing, separate stats. I wish we could use an algorithm as if it were an ETF from within another algorithm. I wish there was an easier way to re-use code between algorithms than copy-pasting. I wish there was a code repository where you could easily grab often-used features or templates, etc. I wish we could quickly compare to multiple benchmarks.
  11. I'll keep dabbling trying to find alpha and backtest -- most of what I find will probably not be Q Fund-able, but the way I see it, a market premise can be utilized on any platform. First and foremost I need to figure out how to get live-trading with a promising algorithm I have that is not yet Q Open material. I have no idea what platform I'll go to. I might check out QuantConnect, or I might try rolling my own that better fits my needs.
  12. I want to make some money. Ideally get to a point where I have a passive income stream that's not sensitive to market downturns.
  13. I read an article about Robinhood and decided on a whim to try it. Then I thought I could write some scripts to automate my trades instead of wasting my time following price movements and placing orders, and a google search brought me to Quantopian.
  1. Brokerage: Robinhood
  2. Live-trading about 4 months
  3. Been on Quantopian about a year
  4. 26k to not be affected by day trader rules. Was about to double it, then came the news.
  5. Returns are good, could be better
  6. Mediocre but sufficient market knowledge
  7. Pretty decent developer. Best in Java, mediocre in Python
  8. Momentum and mean reversion based
  9. Daily and minutely price information. Trying to incorporate more fundamentals and sentiment data
  10. Option trading, more memory, ability to run Keras and Tensorflow for more unsupervised goodness
  11. I'll trade manually. Drop all experimentation on stocks, and continue doing only ETFs and ETNs
  12. Create passive income stream.
  13. Automate trades to remove emotions

What broker were you using?
robinhood
How long have you been live-trading?
1 year
How long have you been working with the Quantopian platform?
1.5 years
How much capital did you invest in your algo(s)?
started with 100 and added until I got to 25k
What kinds of returns did you see? Were they as strong as you expected?
20-30% hard to tell because taking money in and out of Robinhood or Quantopian screws up the %
How deep is your finance/stock market background?
none, self taught
How deep is your programming background? How comfortable are you with Python? What is your "native" or favorite language?
none, self taught
What Quantopian features were utilized to inform your trading signals? OHCL bars? Fundamental data? Alternative data? Daily or minutely? Pipeline?
pipeline, morningstar data
What do you think you'll do now in order to keep live-trading?
zipline live

1. Broker
Robinhood for my taxable account, 7 of my friends/family also using variants of my algorithms in Robinhood
IB for 3 IRAs (one for me, one for my wife, and one in my dad's name)
2. How long live trading
Coming up on a year in November
3. How long Quantopian platform
Coming up on a year in October
4. How much capital
Between all the accounts about $200K
5. What kind of returns
I started with robo-investing algorithms (simple tax loss harvesting and automated order placing between a basket of ETFs). It did as you'd expect for that, everything has done well in the past year. But then I started to be intrigued with what I could do by having a look at investment factors and started playing around with pipeline. At first the stock selection started going on a tear so I put in more money into it then it collapsed and lost all the gains, it recovered most of those losses (was down ~1%). At that point I went back to the drawing board and looked to build a more de-risked multi-factor "index" of 16 stocks across all sectors. I adjusted the factor weighting with a crude machine learning approach by looking at recent performance of factor ETFs and finding an optimal combination of those ETFs - aka an optimal factor weighting scheme. Rebalance monthly. It's a similar approach to what is now submitted in the contest under my Quantopian assigned name: "Bright Orange Cheetah." I only started this approach a month ago and its one I have a lot of confidence in.
6. Finance/stock market background
I'm probably more well-versed than most, come from a Bogleheads background (https://www.bogleheads.org/forum/index.php), read all the classics in finance and investing such as the Intelligent Investor, Four Pillars of Investing, Random Walk Down Wall Street and a few other passive investing books.
7. Programming background
I'm a mechanical engineer by trade but one who really enjoys data analysis. I have worked a lot in MATLAB of recent but did some Java in high school. Python was new but it was/is fairly intuitive. I enjoyed learning the language and that learning will pay dividends for me outside of Quantopian in my engineering career.
8. High-level description of strategies
Started with static asset allocation but some smarts around tax loss harvesting and keeping balances in check (see post: https://www.quantopian.com/posts/why-use-wealthfront-when-you-can-use-quantopian-as-the-robo-investor). This progressed into a dynamic efficient asset allocation model that is finding a trailing optimal efficient combination of assets. This progressed into looking at a dynamic efficient combination of traditional Farma French factors.
9. Quantopian features being used
Daily data and pipeline and the portfolio optimizer, using pysch signal for my market neutral contest entries too
10. What to do now?
Most of my strategies are monthly so I can still replicate this manually. The tax loss harvesting and some other pieces will be a bit more difficult to do manually. I'm hoping/expecting that something will fall out of this after the dust settles and there will be an alternative method.
11. Goals for algorithmic trading?
Remove the emotion from investing. Try and reduce volatility and improve risk adjusted returns. Not necessarily trying to beat the market in the short term, but hope to in the long run. More importantly I'm just looking to have good risk adjusted returns, with little costs, keep engaged in my strategies, and stay invested at all times.
12. What brought you to Quantopian?
Live trading...

  1. Broker
    IB. I have 5 linked accounts by IB. On each one there is a live algo runnning.
  2. How long live trading
    Coming up on a year in November
  3. How long Quantopian platform
    Coming up on a year in August
  4. How much capital
    Between all the accounts about $50K
  5. What kind of returns
    10% to 20%
  6. Finance/stock market background
    Self-learning
  7. Programming background
    I am a programmer. I do alot with Java. Formerly C/C++. But I find Python maybe the best language for handling financial datas.
  8. High-level description of strategies
    Trend-follow, linear regression, Put/Call Ratio
  9. Quantopian features being used
    I fid hte pipeline and the buil-in and customfactor very powerful.
  10. What to do now?
    My statergies are based on daily datas.
  11. Goals for algorithmic trading?
    Simply to say: less drawdown, more returns.
  12. What brought you to Quantopian?
    Live trading and the powerful Python platform.

The most important thing I am facing is to chose an alternative a.s.a.p. But since there are so many choices and I have not so much time to test one after the other, I hope there is someone who can share his experience.

I've been using IB API directly from 2012 till 2016 when I fully switched to Q. In 2015 (I think) IB has finally implemented its API in Python (after 5+ years of requests). Before that I used swigibpy to connect and trade automatically. There is a learning curve in using IB API, and it is, in my opinion, pretty steep. API is pretty much low level. You can learn a lot, but it requires some time (1-2 months max). The downside is that you have to control the connection all the time. So every day I used to turn on my computer, connect to IB, run the algo, and schedule the computer to shut down after market close. This could be deployed to an amazon server, but I just switched to Q.
I haven't used zipline-live yet, but I think its pretty raw and unpolished at the moment. It is in alpha stage. I cannot say that my own scripts are any better, but its my code so don't get lost in it. IBridgePy also seems pretty immature, but I think I will try it. QuantConnect - I don't think about using it for now. It also implies a learning curve, so why not use IB API directly then?

Hi Dias,

I have a post here for disccusing alternative.
https://www.quantopian.com/posts/phasing-out-broker-integration-copy

  1. What broker were you using?
    Robinhood
  2. How long have you been live-trading?
    About a month
  3. How long have you been working with the Quantopian platform?
    About a month
  4. How much capital did you invest in your algo(s)?
    "he's just a poor boy from a poor family" - What I was willing to lose if I totally blew it, a few grand
  5. What kinds of returns did you see? Were they as strong as you expected?
    10-15%, waaaay above my expectations
  6. How deep is your finance/stock market background?
    See Viridian's answer. I have other investment vehicles, but they're mostly fairly passive things like a 401k, so it didn't require a lot of knowledge
  7. How deep is your programming background? How comfortable are you with Python? What is your "native" or favorite language?
    Fairly good, I have a background in engineering and I do a fair bit of programming at work. Python was relatively new to me, but I was actually learning some at work for a different project at the same time. I don't think I have a preferred, I do actually like Python in that it rewards good programming practice and punishes the wicked who can't (or won't) indent properly
  8. Can you give a high-level description of your strategy/ies? Economic premise? Methods utilized? Amount and type of securities used?
    I started off with some Charles Witt VIX stuff, and have been tweaking that. IMHO, the U.S. is going to be fairly volatile for at least another 3.5 years ;-), so volatility-based algos seemed like a good idea. High risk, but I wasn't putting too much money into it and again I do have safer investments elsewhere
  9. What Quantopian features were utilized to inform your trading signals? OHCL bars? Fundamental data? Alternative data? Daily or minutely? Pipeline?
    Minutely and Daily data and OHCL bars, mostly
  10. Were there any features missing from Quantopian that you wish you'd had access to?
    Not really, for me, it was mostly just interface stuff like code respositories, deposits/withdrawals making returns look weird, etc Not anything that really affected my ability to accomplish tasks, just stuff that annoyed me. Basically everything Viridian pointed out was something that I had noticed at some point but just dealt with
  11. What do you think you'll do now in order to keep live-trading?
    Good. Freaking. Question. I have no idea. I don't really want to pay QuantConnect, and zipline-live seems a bit like a newborn infant so I'm hesitant to use it. With more time and such I could probably develop a personal solution and really dig into it, but I have a job and it's a very busy season, so this is something I'm going to have to think about over the next few weeks
  12. What are you goals? What are you hoping to get out of algorithmic trading?
    Make money? Ideally, as I got more comfortable with writing algos and had a community to bounce ideas from, I'd slowly increase my level of investment with the goal of generating a steady revenue stream. I like my job, I don't intend to leave it to do this or live off of investment income anytime soon, but extra cash flow and learning about new topics like this are always great
  13. What brought you to Quantopian?
    I was looking into Robinhood and read about Quantopian on Reddit. 5 minutes later I was coding an algo
  1. What broker were you using? IB, 6 accounts, 5 linked, 1 seperate with trust fund money
    1. How long have you been live-trading? with Quantopian mid 2015 so 2 years? Via other means 30 yrs
    2. How long have you been working with the Quantopian platform? 2 years
    3. How much capital did you invest in your algo(s)? 110K would love to deploy 500K
    4. What kinds of returns did you see? Were they as strong as you expected? -5% to 25% p/a (DD < 6%). As expected In guess. Some algo's performed too different from paper so I turned the off when they deviated: that was a dissapointment
    5. How deep is your finance/stock market background? Quite deep: trading for 30 years, MBA, statistical background through MA sociology. I trade options, futures, stocks, etfs. no bonds and CFDs
    6. How deep is your programming background? How comfortable are you with Python? What is your "native" or favorite language? Programming OK, but no superstar material in terms of software design. I like Python, I'm OK in Perl, medium in C type languages
    7. Can you give a high-level description of your strategy/ies? Economic premise? Methods utilized? Amount and type of securities used? I have three types of strategies: a) I have a few low-risk algo's that trade momentum and trend reversion like VAAG; b) I'm a VIX future junkie, and I get by in Quantopian by trading XIV and UVXY but would prefer trade options and futures. I allocate up to 10K to a VIX algo: these are my barbells (antifragile theory); c) the final type is pipeline algo's where I scan for fundamentals and third party datasets and try to find the winners. I would love to add long-short stuff, but IB in Oz doesn't allow that (need to connect to another broker)
    8. What Quantopian features were utilized to inform your trading signals? OHCL bars? Fundamental data? Alternative data? Daily or minutely? Pipeline? All of them, Price mostly minutely (for stoploss detection), rest daily. I use trading calendar, sentdex, psychsignal.
    9. Were there any features missing from Quantopian that you wish you'd had access to? Would love COT data, minutely fetcher, minutely Indexes(important), ability to email actions, or json it in addition to logging and recording. Record more than one custom graph with more then 5 indicators, download all transactions, send orders to more than 1 account
    10. What do you think you'll do now in order to keep live-trading? Experiment with IbridgePY, look at QountConnect, but I assume I'll have to find a cloud platform I can work with or spin up an AWS instance of something
    11. What are you goals? What are you hoping to get out of algorithmic trading? More money, no money to money managers, earlier pension: maybe even start a company trading for others.
    12. What brought you to Quantopian? I started with RIZM, now defunct, and I continues with live trading via excel (yes really with VB) and could not get it stable, then I discovered Quantopian. If Quantiopian would not have live trading
  1. What broker were you using?
    IB
  2. How long have you been live-trading?
    Since 2012. Before Quantopian I used the IB API directly. Connected from python.
  3. How long have you been working with the Quantopian platform?
    Almost 2 years.
  4. How much capital did you invest in your algo(s)?
    350K
  5. What kinds of returns did you see? Were they as strong as you expected?
    In Quantopian: from -7-8% to +20%. In general: from -17% to +35%.
  6. How deep is your finance/stock market background?
    6 years in a pension asset management firm on various positions from analyst to portfolio manager, 2 years in a sovereign wealth fund as a risk manager. CFA charterholder since 2012.
  7. How deep is your programming background? How comfortable are you with Python? What is your "native" or favorite language?
    I've been programming Python since 2012. Started with writing a program for trading through IB API actually. Python is my preferred language, feel very comfortable with it. I used at work as well in my role as a risk manager. I cannot say that my knowledge is deep, but its ok for implementing algos and financial modeling.
  8. Can you give a high-level description of your strategy/ies? Economic premise? Methods utilized? Amount and type of securities used?
    The current strategy is basically my own asset allocation approach based on risk targeting used together with volatility-based momentum indicators to define long and short positions. The premise is that asset volatility and correlations between asset classes are both a risk and momentum indicators. The main instruments are asset class etfs. Other strategies are in development at the moment. It takes 6-12 months for me to develop a strategy for live trading.
  9. What Quantopian features were utilized to inform your trading signals? OHCL bars? Fundamental data? Alternative data? Daily or minutely? Pipeline?
    Basically historical pricing is the only thing that is needed. Minute-level data is employed. Pipeline data is not used, because the algo is pretty heavy even for few instruments, cannot run it over hundred assets. I use ipython notebooks a lot.
  10. Were there any features missing from Quantopian that you wish you'd had access to?
    The speed of backtesting is far from best, but is acceptable. This makes you re-write the code more effectively. Real-time index data would be a huge plus.
  11. What do you think you'll do now in order to keep live-trading?
    Probably , I will return to direct connection using IB API. But I will look at the development of the zipline-live. Very skeptical about QuantConnect.
  12. What are you goals? What are you hoping to get out of algorithmic trading?
    Honestly, I actually put my life on it. I don't know if it is a right bet. At the moment this is all I do, so I try to work hard and learn what I can. Time will show.
  13. What brought you to Quantopian?
    Actually, I wanted to develop a backtester myself, but life circumstances actually made me focus on developing strategies, not programs. Q was a solution plus the convenience of deploying the algo for live trading with a few clicks. My algos didn't meet the contest criteria, they are more focused on squeezing beta returns rather than capturing alpha from some factor.

1. What broker were you using?

Robinhood.

2. How long have you been live-trading?

1-month

3. How long have you been working with the Quantopian platform?

less than 2-months

4. How much capital did you invest in your algo(s)?

$ 5k and moving up next to $25k –

5. What kinds of returns did you see? Were they as strong as you expected?

Reasonable alpha and risk management depending on strategy

6. How deep is your finance/stock market background?

moderate investing background and experience in asset allocation

7. How deep is your programming background? How comfortable are you with Python? What is your "native" or favorite language?

I am liking python. It is a new scripting language. Current background is engineering with scripting and numerical analysis.

8. Can you give a high-level description of your strategy/ies? Economic premise? Methods utilized? Amount and type of securities used?

first looking at mean reversion - prior to exploring trading at cost data sets. I was interested in the development of long short but the platform changes have resulted in a full stop.

I was still in the process of reviewing the tax implications expanding trading. The live / paper trading was the single most informative feature to understand the market day to day.

9. What Quantopian features were utilized to inform your trading signals? OHCL bars? Fundamental data? Alternative data? Daily or minutely? Pipeline?

mostly pipeline and fundamental data

10. Were there any features missing from Quantopian that you wish you'd had access to?

access to data feeds or external files in real time. Would have allowed for development of scripts. Also wish there was more realtime trading / paper trading information on performance.

11. What do you think you'll do now in order to keep live-trading?

I don't know yet. I need to find a replacement to Quantopian to support trading and exploration. I had hope to contribute to Quantopian. I really have gotten a good picture of the community from this live trading debacle. I hope that the community can continue in some way.

12. What are you goals? What are you hoping to get out of algorithmic trading?

I would consider algorithmic trading as a career change. I have been inspired by the programming and data analysis. The community has been great for that. Like others I'll need to implement another solution to realize goals.

13. What brought you to Quantopian?

It was by luck. Looking for assets allocation optimization and the community and tool was amazing.

What broker were you using? Robinhood. I pulled my money from my account today, and will likely be going to IB or just quitting altogether.

How long have you been live-trading? Since May 30th.

How long have you been working with the Quantopian platform? Early/mid-May.

How much capital did you invest in your algo(s)? I had up to 30k in my Robinhood account, but I generally only allowed it to play with the amounts in excess of $25,000.

What kinds of returns did you see? Were they as strong as you expected? Hard to say. Day one saw a 6% gain, which was wiped out on day 2 solely because I made some poorly thought out changes overnight. Since then I've only deployed sparingly, with one day of a 2% or so gain, and a few other days where either no signals were triggered, or the gain/loss was negligible.

How deep is your finance/stock market background? Uh, five, six months worth of experience? I just dumped whatever I could in my 401k, set it to the most aggressive level, and didn't think much of investing until I had enough money sitting around doing nothing that I decided to play with it. I tried a couple of roboinvestors but wasn't interested in them, and my first two months on Robinhood were a disaster because I didn't know what I was doing.

How deep is your programming background? How comfortable are you with Python? What is your "native" or favorite language? Aside from learning QBASIC in high school and teaching myself basic HTML around that same time (which I have since forgot), I didn't know anything about programming or Python until I came across this site in early May. I took a week off of work to learn Python and technical analysis, and worked at it until I cobbled something together that A) worked and B) I understood why it worked. The funny thing is, after glancing at QuantConnect's site, I'm like "wow, whatever this is is so different than what I know, I don't know if I'll ever get comfortable with it", even though three months ago, I didn't even know what Python was.

Can you give a high-level description of your strategy/ies? Economic premise? Methods utilized? Amount and type of securities used? The theory behind my strategies was that if I have the 25k to day trade, and it doesn't cost anything to make a bunch of moves, that I'd try to find momentum in relatively liquid, higher risk securities (typically 3X ETFs), get in for a small profit, and get out with minimal risk. Disregarding the tax benefits of buy and hold, vs. scalping, you can beat the returns on any stock with a long-only strategy if there is no cost to trading.

What Quantopian features were utilized to inform your trading signals? OHCL bars? Fundamental data? Alternative data? Daily or minutely? Pipeline?
Were there any features missing from Quantopian that you wish you'd had access to? Basically just OHCL and daily price data. Pipeline would have been much more useful for me if I was on IB, since I'd be able to go short. Even filtering out penny stocks, finding big gainers from the previous day and shorting those is seemingly a no-brainer. I wrote (and unfortunately deleted) one strategy that did exactly that, and while I don't think I ever let it run to completion, the results would have been astronomical. I don't think I even wrote in any technical indicators, it just said "if pipeline says this gained a lot the previous day, short it and cover by EOD."

What do you think you'll do now in order to keep live-trading? Maybe QuantConnect, maybe look at what else is available in IB, or maybe just switch over to cryptocurrency. I can't go back to manual trading, it's too difficult to do while I'm trying to also work a full time job, and I have too much of a trigger finger when it comes to timing buys and sells. The strategies I wrote were much more disciplined.

What are you goals? What are you hoping to get out of algorithmic trading? I want a race car.

What brought you to Quantopian? Luck, I guess. I was looking for Robinhood integrations so I didn't have to make trades from my phone, and the only ones I found were this and StockTwats. I needed an education and it was clear that I'd get more from learning code and reading the forums here than I would from reading tweets like "HAY GUYZ $DRYS IS GOING TO THE MOON, COME OVER QUICK".