Yield Profiles (Markouts) - Overview
Yield Profile Contents
Legacy Open/Close PnL
Historically broker's have looked at whether a client is systematically making money from trading by observing their month to month open/close PnL i.e. the money the client makes from closing their open positions.
This whilst useful, can of course occur only after they have successfully closed a lot of profitable (for the client positions).
The current value of realised + unrealised PnL the client has made, whilst useful can be subject to a fair amount of PnL variation over long holding periods.
Mahi uses an analysis called Yield Profiles or Markouts as an alternate method of assessing client flow. It is a good indicator of the quality of client's flow. This earlier indication of flow quality can spot bad client's earlier and reduce the time to action
Yield Profiles / Markouts
Yield profiles compare the trade price to the market mid price at various intervals after a trade is executed.
We produce these for each trade and look at how the flow stacks up in aggregate as if all trades were done simultaneously.
It is harder to predict where the market will go the further away from when the deal was done.
Latency arbitrage and bad client flow can be observed by immediate bad impact to client mark outs.
Through grouping (the field by which you aggregate analysis by) and filtering (the field by which you constrain analysis by), the Yield Profile section of MFXEcho is a tool to drill down to a granular level of detail.
What is a Yield Profile?
Yield Profiles show the value of a trade, counterparty or instrument from the house perspective, so a positive yield profile reflects an increase in PnL for the house rather than the trader. They are an indicator of profitable flow, harmful counterparties and optimal holding periods.
Aggregating yield profiles allows for higher level analysis, as MFXEcho constructs a yield profile from the entirety of the underlying filtered data. This provides insight into the overall quality of the underlying trade flow.
How is it Calculated?
A yield profile is calculated for each trade that MFXEcho receives via a process called risk path.
The yield profile is plotted by subtracting a Mahi generated mid price from the trade price. This is repeated at various time intervals to plot how the trade preformed against the market.
The spread generated can be seen by reviewing the yield at 0s time horizon.
The time at which the trade has cost the house is found by looking for the time where the trade goes below 0 yield on the y axis.
What is it used for?
It is possible to find trends in data by viewing the aggregation of yield profiles, then breaking down further by grouping and filtering data. MFXEcho can generate yield profiles for multiple different fields, including book, instrument, counterparty and more. Filtering and grouping can be used to drill down and identify subsets of flow behaving in a similar way or exposing the most damaging accounts. For example filtering by instrument may expose weakness in a specific instrument's pricing, for example, during a particular timezone.
Historic trade data is held in the MFXEcho database, meaning yield profiles can be analysed for hourly, daily, weekly or monthly periods. This is particularly useful when analysing individual counterparties as it helps determine whether an account with a negative yield was simply lucky one week or is consistently costly to the book over the longer-term.
Back to Flow Analysis Contents