Proof Trading Progress Update: 15 Months

It feels strange to be writing this quarterly update with everything that’s going on in the world, but we do want to maintain our regular cadence.

As a company, Proof hasn’t been directly affected by the Covid-19 crisis, but all of us on the team are shaken. We’re worried about each other, our families, and our communities, as well as the potential long-term impact to the economy. We wish there was more we could do to help. We’re working from home and trying not to be a transmission vector. As the markets have gone crazy, we’ve cut back on customer outreach and are just trying to stay out of people’s hair.

If we were already live, we’d likely be fighting fires right now, and if not maybe we’d be urgently conducting research to try to enhance/correct any sub-optimal trading behavior in these erratic markets. Since we’re not live yet, we have the benefit of time. We can systematically and thoughtfully examine current market conditions in comparison to “normal” market conditions and incorporate all of our findings into our day one product. Unfortunately, we’re just not yet at a point where we can confidently produce timely and actionable research that we think would be helpful right now. And what we’re not going to do is try to capitalize on the moment by churning out half-baked click bait to capture mind share and market share. So we’re just sitting on the sidelines, chipping away at the mountain of work in front of us.

We are making progress toward launch, although we have taken our foot off the gas pedal a bit these past few weeks as we prioritize family and well-being over work. We feel very fortunate that we have the luxury to do so, and for a number of other reasons:

• We’re pre-launch, and this is a heads-down period anyway while we finalize our research and development. So far at least, we haven’t seen any material delays to our launch timeline (Q3 2020).
• Proof has been operating as a distributed team for over a year, so there hasn’t been much disruption to our day-to-day.
• We’re well capitalized and we have no plans to raise additional VC funding. At our current burn-rate, we have over 5 years of runway. Even assuming a full production ramp up in July and zero revenue, we have approximately 18 months of runway.

We hope everyone is staying safe and sane out there, and as always if there is anything we can help with, please let us know.

Major updates

• Proof Services LLC, our broker-dealer subsidiary, submitted Form NMA to FINRA on January 19th.
• Made substantial progress on the product and technology fronts.
• Published an in-depth account of how we selected our clearing firm and submitted a comment letter to the SEC recommending changes to the SIPs.

Goals from our previous update

Technology

• Continue the build-out of the business components of the trading system such as the FIX Gateways (client-facing and venue-facing), the Market Data System, the Order Management System, and the Algorithmic Trading Engine.

(6/10) We’ve made progress on all of the above-mentioned business components, but work still remains to integrate the numerous pieces and tie up loose ends.

• Select an extranet provider and establish our venue access strategy for lit as well as dark venues.

(9/10) We’ve selected an extranet provider and have nailed down most aspects of our connectivity to customers, venues, and market data.

Regulatory

• Proof Services should file Form NMA and be extremely responsive to any follow-up requests from FINRA.

(9/10) Proof Services submitted Form NMA on January 19th and has already received and responded to its first follow-up request from FINRA.

Product/Research

• Select a base feature set for our historical simulation approach and measure its predictive strength.

(5/10) We selected a base feature set that is showing non-trivial predictive strength, but we don’t yet know if it is strong enough to drive improvements in algo design.

• Craft our first iteration of execution metrics and present the full process of their design.

(9/10) This is internally completed but not yet fully written up for external release.

• Design the first iteration of our core trading algorithm

(2/10) We’ve started on this process, but we have not yet derived a design for the first iteration.

Sales

• Continued focus on buy side engagement and producing content.

(4/10) We did produce some content this quarter, though not as much as usual. Our plan had been to do a sweep of our buy side contacts to get feedback on our execution metric research and our day one plan for our trading algorithms. With everything going on, we’re continuing to perform this research on our own, but we’re trying to stay out of people’s hair until things hopefully calm down. If you’re reading this and on the buy side, and you do have capacity to let us pick your brain, please let us know!

Operations

• Begin integration with approximately 5 major EMS providers and be on track for production-readiness at launch.

(3/10) We held “integration” conversations with two major EMS provider in January and February. Based on these two conversations, it seems there will be a wide range in our experience integrating with EMS’s. One provider said the process should take 2 weeks start-to-finish and can begin as soon as we lock down our fix spec, but the other cautioned that they can’t even begin the process until after we receive B/D approval and that it will likely take at least 1–2 months. We hope to get introductions from prospective customers to additional EMS providers if and when things calm down.

Goals for the next 3 months

Technology

• Development complete, including QA testing.
• Begin production environment design and deployment.
• Make significant progress on all external integrations.

Regulatory

• Proof Services should continue to be extremely responsive to follow-up requests from FINRA, particularly as the end of 180-day approval timeline nears (i.e. late July).

Product/Research

• Lock down and implement the day one design of our core trading algorithm(s).
• Design sanity checks and performance testing for algorithm designs using historical data (to the extent possible).
• Design a framework for A-B testing algo designs that can be used once we are trading live.

Sales

• Identify day one (i.e. “founding”) customers.
• Continued focus on buy side engagement and producing content.

Operations

• Determine partner for DMA and set timeline for integration (e.g. pre vs. post launch).
• Have a clear integration plan/timeline for the respective EMS provider of each day one customer.
• Establish a formal process for evaluating potential new hires.

Here are our full updates across the board:

Technology

• We have partnered with an extranet provider and designed a connectivity solution that allows us to run our trading system in the cloud (AWS) and access counterparties located in various data centers in NJ and Chicago.
• Business components: (a) We have completed the development of an Exegy-based market data service and are working on integrating it into our trading platform. We have also signed agreements with CTA and UTP to be able to receive SIP feeds immediately. (b) We have created a resilient and high-available version of QuickFIX/J and incorporated it into our FIX gateways. (c) We have completed the design of a full-featured OMS component that will significantly simplify the design of the algorithmic trading engine. (d) We have made significant progress in the design and development of a UX application framework, including a web-based real-time trade blotter component.

Regulatory

• Proof Services submitted Form NMA to FINRA on January 19th. This is the main application for becoming a broker-dealer.
• Proof Services received its first request for information from FINRA on February 11th and submitted a response on March 17th.

Product/Research

• Completed our first iteration of execution metric design.
• Began design of the core trading algorithm(s).
• Identified a promising candidate feature set for historical market simulation.

Sales

• Our main initiative on the sales front continues to be performing execution analysis for the buy-side. We have received an exceptional set of trading data, and we spent a large portion of this past quarter analyzing it. We expect this research will continue to be a high priority for the next several months. This project has proven extremely helpful as we homed in on our first iteration of target execution quality metrics, and we expect it will be equally useful in informing the design of our execution algorithms.

Operations

• We do not plan to directly connect everywhere out of the gate, and instead we plan to access most exchanges and dark pools through another broker. There are two main reasons for doing so: 1) there is substantial overhead in establishing and managing venue connections and 2) almost all exchanges have volume-based pricing tiers and they generally provide much better pricing to members who hit those tiers. In many cases it is more expensive to access those venues directly than through a larger broker’s MPID to take advantage of their tiers. We had conversations with four potential DMA providers, multiple of which seem promising.