When do DARWINs close for new investment?

With several DARWINs gaining Assets under Management at speed, the time has come to announce the DARWIN lock-out methodology.

This blog post explains how the process is monitored, with two new areas in the investor interface providing the relevant information.

Reminder – Divergence

DARWINs track their underlying strategy. DARWIN provider and investors are bound NOT to trade at the exact same prize because:

  1. Investors’ replica block trade lags the lead DARWIN provider’s
  2. For popular DARWINs, the investor block trade volume exceeds provider’s

The market typically moves during the latency period, and this introduces a performance difference. On some trades investors lose out, on some they win. Over a long enough sample of trades the latency effect averages out – but it introduces a bit of noise in the replication process.

Unlike the latency effect, the volume effect makes investors systematically worse off. They lose out gradually as growing trade sizes (driven by growing assets under management) wipe available market depth. For successful enough DARWINs, there comes a point when a marginal EUR / USD / GBP invested yields no marginal profit. This point marks a strategy’s AuM ceiling.

Monthly Divergence

The single metric driving when and how a DARWIN is closed for new investment is monthly investor divergence.

Monthly investor divergence is public for all DARWINs on the public dashboard, with a high-level summary (A in the screen capture below) and a detailed overview (B in the capture):

Screen Shot 2016-06-16 at 17.47.48

Negative monthly divergence (red) indicates how much DARWIN investor performance lags DARWIN notional performance, per month. Note that to protect investors from this effect, investor trades clear at the raw spread provided by our LPs. This is the reason why most DARWINs sport positive (green) divergence by default. The monthly divergence  figure is calculated for the most representative trades by the strategy (note – there are numerous hedging micro-trades triggered by the risk management algos to “shave” investor exposure, but these do not count towards the divergence computation).

The public submenu in the DARWIN profile (B in the screenshot) includes two tabs.

The first tracks divergence in cumulative % – as a difference between DARWIN and investor performance:

Screen Shot 2016-06-16 at 20.27.00

The second tracks pip divergence: tracking positive (green) and negative (red) divergence accounting for replication latency (average and median) in milliseconds (ms) and individual trade size (bubble size): Screen Shot 2016-06-16 at 20.29.38

Note that divergence can fluctuate substantially across strategies and market times. It can be particularly slow during market roll-over and news releases, as trades take longer to fill and the execution bridge takes longer to report to our investor infrastructure.


What’s next?

As AuM for the more popular DARWINs scale, we’ll continue to monitor all the parameters driving investor divergence, with a view towards:

  1. Refining the calibration of the scalability score to AuM
  2. Developing algorithms that fraction investor execution – it is certainly possible to reduce slippage this way
  3. Offering providers toolkit to optimize their investors’ liquidity trade-offs
  4. Refining the threshold after which DARWINs are closed for new investment (currently set at -1% negative divergence)

Thoughts? Suggestions? Go ahead!

0 replies

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

Your email address will not be published. Campos obligatorios *