I posted the discussion Automate Successful Manual Trading Strategies on linked-in where I have over 3,000 followers hoping for a more lively discussion. Please add your comments at linked in where we will have the full stream.
The essence of the question:
is it good idea to apply our skills and existing trading code and use them to automate manual strategies other people who are trading successfully and get paid via a share of the profit from the automated trading?
It seems that many traders with 3 to 10 million of capital traded by manual strategies may not be able to afford the risk of automating the strategy. They will also be hesitant to incur an unbounded maintenance overhead. We can reduce labor costs by leveraging our code base of existing automated strategies and associated plumbing and connectivity. In addition we are experts in exactly this kind of complex automation and really enjoy complex difficult problems and find trading algorithms absolutely fascinating. I like business models with profit tied to the results of our activity rather than labor expended.
I have not committed to this approach but do see compelling aspects that seem to fit our strengths and the way I would like to see the company positioned in the market. I am very interested in feedback.
- I can see several risks we may implement strategies and never get paid because
- Something was missing in the spec that turns out too expensive to automate.
- The strategy contains a fuzzy aspect we can not automate.
- The strategy really isn’t profitable but this was masked during analysis.
- The customer / partner ends up not trading
- Because they don’t really have the authority to place the capital at risk.
- They decide the profit share is too expensive.
- They decide to attempt automation using internal resources.
- They fail to deploy the hardware needed to support the software.
- They can not get their broker to allow API access to their servers.
- They don’t have access to requisite data in automated form.
On the other hand we will be working to automate trading strategies anyway and we may learn something from other experts along the way. I am strongly motivated to develop recurring long term revenue options that are not tied to Time and materials consulting.
Trader / Partner / Strategy Criteria
I suspect our first group of traders will be independents managing a total between 3 and 60 million USD. This group would have a hard time funding internal R&D / development staffs but could have some wicked smart manual traders active in their company. In the first generation we would want strategies that are using mostly pricing data and technical indicators.
- Customer is trading at least 2 million USD in a manual strategy.
- Strategy has annual yield in excess of 15% of capital.
- Strategy has been profitably trading live for more than 3 years .
- Customer will start trading automated version with at least 2 million USD once tested.
- Customer has desire to automate the strategy.
- Customer has authority to agree to profit share.
- Customer has authority to share strategy details.
- Customer understands strategy well enough for us to automate it.
- Customer has legal right to use the strategy.
- Customer has or can obtain access to automated API at broker.
- Customer has or can obtain access to all required data in automated fashion.
- Strategy will be essentially the same when automated.
- Strategy does not require subjective human input that can not be automated.
- Customer is likely to continue the relationship if automation is successful.
- Customer has additional strategies with additional capital we can automate if mutually agreed.
- How many traders have existing manual strategies trading enough capital?
- How many of these traders will be interested in automating?
- Why have they not already automated these strategies?
- How do we confirm they are ready / willing and able to execute?
- Will the profit share be adequate to justify our costs and Risk?
- What is a rational profit share %?
- Is there a middle road where we share a portion of the risk for lower % of profit share?
- What factors should be used to adjust the profit share %?
- What will it cost us to separate automatable strategies from the fuzzy unconstrained set?
- What will it cost us to update the strategies to reflect changing market conditions?
- How much of this can be parameterized so it can be handled by the customer?
- Will the customer do the work to update the parameters to re-optimize?
- What tools will they require?
- Who are the best set of customer / partners
- Where do they work?
- What are their Titles?
- How would they get this authorized?
- Who has to participate in the decision making process?
- How do we find and contact traders with the right combination of capital and strategies?
- What blogs and magazines do they read?
- What trade-shows do they attend?
- How do we identify the best customers?
- How will the customers / partner set evolve over time?
- Who else is using the same approach? An how has it worked out for them?
- What if any regulatory codes will apply?
- Why should we not take this approach?
- What is the minimum acceptable capital trading in the manual strategy?
- What is the minimum acceptable yield in the manual strategy?
- What is the minimum acceptable history from live trading the manual strategy?
- Is there a class of manual strategies we can make more profitable with automation?
- Should we attempt to do this since it seems more risky and more time consuming?
- How would we identify the subset which offer highest profit share potential?
- How do we isolate fuzzy aspects of manual strategies so we can determine which fuzzy aspects could be automated from those which truly require human intelligence?
- What kind of integration with existing customer IT infrastructure will be required.
- How will this change between customers?
- How will this integrate with their Audit and Risk control department?
- The first 6 months:
- What marketing channels should we use?
- What marketing companies should we engage?
- Which sales people should we engage?. Company? or Characteristics?
- What proof points will customer require?
- What marketing assets do we need to develop?
- What will average customer require to validate the Automated strategy?
- How long with this process take?
- What tools do we need to include to support validation?
- How to we formalize this so we don’t spend too long in validation?
An interesting client set would be senior or partner level traders at hedge funds. A good hedge fund may have completed some automation but is backlogged for requests to automate strategies. I have seen this phenomenon in other businesses where executives have needs or strategic goals their internal IT can not satisfy. If we could tap into this market it seems like a win/win. I don’t know if they could obtain permission to share the information we would need to automate the strategy but the enlightened CIO, CTO or CEO should see us as a low risk way to move forward on strategies that could be profitable but are too small to gain priority on their product management backlog.
Thanks Joseph Ellsworth