How can we help?
Ed Howorka Solutions company was founded by me, Ed Howorka, to make a living doing what I like to do most: helping others to solve their problems.
I am what you would call a multi-disciplinarian problem solver. My career started with a fascination and formal training in theoretical math and physics. It progressed through many phases, including computer science, AI, robotics, signal processing, trading systems and interfaces, large scale communication networks — to list just a few. My current passion, that consumed quite a few years of ardent studies and effort, is the field of automated trading strategies, a vast and challenging subject that uses everything that I ever learned and (infinitely) more.
My consulting services are mostly in the areas of Forex trading strategy evaluation, development, historical back-testing and optimization. In addition, I welcome questions and assignments in any other trading technology area. If deep expertise is required outside of my knowledge domain, I can often rely on advice from specialist colleagues. Then — of course — once I agree to an assignment, any effort by me to acquire the necessary knowledge is “free of charge.”
To give you some idea of “real-life” questions and assignments that I’ve dealt with, here are a few examples:
From a head trader…
We have a trader (“Joe”) that is very successful trading manually (as demonstrated by his record over several years). He trades every few minutes, and makes a nice profit before closing all his positions for the day. The objective for us is to scale up his trading style by adapting it to other instruments and by automating it. We have debriefed Joe extensively and documented his trading rules. However, all efforts to code an automated Joe strategy (“AJOE”) based on his rules have failed. Historical analysis shows that AJOE is unprofitable and utterly fails to mimic Joe’s actual trades. It seems that Joe has a sixth sense that plays a bigger role than any rules that he can explain to others. Is there a way to formalize his intuition as an automated strategy based on the historical record of his trades in the context of market data and news feeds?
Deliverables: A documented approach to accomplish the objective, including supporting software.
From an algo trader…
I have developed a rule-based trading strategy based on years of experience trading this instrument manually. My rules are based on conditions and equations that are parametrized with close to a hundred of defining constants. I used a number of software tools trying to find the optimal combinations of parameters. Some runs took several weeks on a very fast PC. The resulting strategy is disappointing – it is profitable on long timescales, but has big drawdowns. Can you suggest a better approach to optimize the strategy and reduce the risk?
Deliverables: An analysis report and software to discover and maintain optimal parameters as gross market characteristics change over time.
From an independent trader…
I have a very profitable rule-based trading strategy (daily Sharpe ratio consistently above 0.4 for the last year; only a few weeks with negative returns). Then, over the last few months, the strategy’s performance deteriorated to the point that it is barely profitable, and now I had my first month with negative returns. Volatility and other gross market characteristics that I looked at are all within the usual range. How do I find the reasons for the bad performance? How do I go about fixing the strategy?
Deliverables: An analysis report suggesting the likely reasons for the strategy’s failure and suggesting ways to repair it.
From a hedge fund analyst…
The excess spread we pay for market orders is correlated to the order execution time – generally, the longer it takes to receive a deal confirmation, the worse the deviation from price we expected on order submission. Should we make an effort to lower order messaging latency?
Deliverable: An answer supported by explanation of venue microstructure characteristics.
From a product manager…
On our trading venue, traders that are more geographically distant (and experience the greatest network latency) are more successful than “nearby” traders in executing aggressive limit orders. What causes this?
Deliverable: An aswer backed by a detailed analysis report.
From a federal regulator…
Looking at several years of transactional market data in this Forex currency pair, we discovered that there are consistently more aggressive buys than sells (per day) in this instrument. What causes this?
Deliverable: A speculative aswer backed by a detailed analysis report.
Consulting engagement process
The first step is to contact us and set up an appointment for the initial consultation. At our discretion, the initial consultation may be by phone, Skype, or face-to-face. The consultation is free of charge.
During the initial consultation you will describe the general problem and we will describe the process that will be followed to solve it. The initial consultation is implicitely covered by our strict Initial Consulting Non-Disclosure Agreement. In particular, we will keep confidential the fact that you contacted us, or the questions or problems that we discussed. On the other hand, there are no restrictions on the use of the information that you may gain. If you wish, we will sign a formal NDA of your choice prior to the initial consultation.
The objectives of the initial consultation are to get to know each other, define the problems that need to be solved, the information that will be required to start working on the solution, and the terms of the assesment phase. The projected timeframes may include the time required for us to get up to speed with specialized topics. We do not charge for this time – you are hiring experts and it’s our responsibility to acquire additional expertise, if necessary, at no cost to you.
This purpose of this step is to understand the underlying details and to define the assignment in terms of deliverables for the solution phase (next). On your part, this means providing access to expert resources, historical data, product information, source code, etc. This phase is usually covered by the same NDA as the initial consultation, which now needs to be formally agreed to, however.
The result of this phase may be the ultimate answer/solution to your problem. But, in most cases, it will lead to a more rigorous formulation and — perhaps — a reformulation of the original problem. Experience shows that even the most daunting problem can be solved more easily once it is stated in the clearest possible and narrowest terms. Accordingly, as part of the initial assessment, we ask a lot of questions and get to understand the devilish minutiae. In some cases, this may involve specialized tools, custom software, outside experts, and studies of related issues. The ultimate goal is to discard the non-problems. To paraphrase Arthur Conan Doyle’s famous quote: “Once you eliminate the inessential, whatever remains, no matter how weird, is the real problem.”
At the end of the assessment phase, we will present our findings to you. If the original problem is not resolved, we will present a proposal for the solution phase including specific objectives and deliverables, estimated timeframes, research, documentation, design, high-level specs for custom code, performance guarantees, and expected costs. It may take a few meetings to arrive at a proposal that meets your expectations.
This phase delivers the ultimate solution to your problem.
The solution phase proposal may require additional legal agreements. Specifically, it may require a new “bilateral” NDA where you agree not to disclose certain solutions to the outside world (with anything known to you beforehand being usually excluded from the NDA). You may (or may not) agree that we will be allowed to reuse the general solution or pieces of the newly developed software in our future consulting work. If the delivered software is based in part on open source code (which may dramatically lower your costs), you will have to agree that these parts of the modified software will be made available under the same license, and/or have the original author’s copyright acknowledged, etc., etc.
The solution fees are agreed on at this point. Some possibilities are: a flat charge, time-billing, or an initial fee combined with a yearly licensing fee if the solution is accepted by you.
Things to keep in mind at this stage:
- We expect to have an easy access to a manager that is authorized to make cost/benefit decisions as the work progresses and that has access to expert resources within your company. This requires frequent regular project status reviews.
- To minimize your costs, you may have to offer dedicated resources. This usually lowers both the development and the (future) training expenses.
- The assessment that we offered in the previous phase may change during the solution phase when new problems are discovered or new business or technical objectives are defined.
- In spite of all the effort spent in the initial assessment phase, the solution phase may end in a failure to find the solution. We emphasise this risk in the assessment report. But, even in this unlikely scenario, the customer (you) still benefits by having a proof that the seemingly promising way was a no-go. (We often tell business managers that losing a design engineer means a great and unrecoverable loss of the memory of failures — not successes — because although failures are remembered, they are never properly documented. In problem solving, knowing the wrong way is often as important as knowing a possible right way.)
This phase is only required if we deliver software or ideas that have lasting values.
You may require us to promptly fix bugs, advise and/or train your staff, re-adapt trading strategies, etc. At this time we are not ready to operate or monitor your systems in real time.