##FX/eq projects ] PWM, for SG IV

—-fx or eq
* Swing app to monitor multiple live database tables and auto-refresh upon new quotes/trades and position updates.
* Post-trade Affirmation for complex derivative deals. A workflow application. Operations verify deal attributes by phone and Affirm on the deal to send out trade confirm.
* Statistical feed “blender” (with multiple purposes), needed to remove outliers among diverse indicative eFX forward quotes. Outliers tend to mess up our auto pricers, tiered pricers and cross rate pricers.
* (Ledger balance?) PnL report. Nightly batch to apply marks on option positions and compute PnL. Reports both clients and firm accounts.
* Volatility surface fitter for FX (& equity index/single-names). Volatility smile curve fitter. Quote inversion. Option position mark-to-market, driven by live quotes from multiple live eFX markets.
* Workflow engine for Error trade reconciliation/chargeback. A single error ticket can compute PnL and chargeback across multiple (up to 100) error trades and cover trades. Every quarter, this engine reconciles hundreds of error trade whose total values add up to millions of dollars.

* Deal entry screen for trade booking after a retail dealer executes over phone. If trade validation fails, we notify the dealer to amend and resubmit. Same swing app is also used by spot/swap/options dealers to book voice trades.
system? data normalization to take care of splits and dividends
system? basic back testing of simple strategies
system? re-sampling

* (EOS) web-based trade/order entry, with multiple validations. Used primarily by investment advisers across Americas, Europe and Asia. Thousands (up to 5-figure) of orders received daily.
** showing real time bid/ask from GS + bid/ask from trading venues
** supports most order types — limit/market/FOK/IOC/SL

* swing-based real time order book display. Updates to displayed order books come from the backend OMS via real time messaging.
* Also contributed to the OMS backend. Orders originate exclusively from private bank clients. We then monitor their cancels/amends and execution reports (via firm-wide messaging hub) from exchanges, ECN and GS private liquidity pool. Total message volume up to 6 figures.

We are the interface to multiple ECNs to 1) receive quotes, enrich and publish to PWM clients 2) forward client orders to ECN in 2-leg spread trades 3) execute trades received from ECN 4) generate and validate (against tri-arb) cross rates using ECN quotes. Also contributed to auto quoters and RFQ engine. Supervised and monitored the progress of high-frequency eFX applications. Performed eFX development activities such as requirement gathering, design, testing and deployment. Personal contributions included
* Tiered quote pricer. Each FX customer falls into one of the tiers. When a currency pair updates in our pricer, all registered clients would get a new quote (??delivered to their Financial Advisors??). Silver tier clients receive a better bid/ask spread than regular clients; Gold tier gets the best quote.
* eFX rate update/distribution engine to downstream real time risk, PnL, position marking systems
* eFX option real time risk report (unfinished). Option position risk calc is rather slow, so each user selects a limited number of positions into his watch-list. Watched positions get periodically updated based on live ECN rates to show real-time risk.
* (questionable project) Compute cross rates for PWM trades that are large, recurring and involving illiquid currencies. Cross rates computation from daily close USD buy/sell rates.
— blender
http://www.olsendata.com/fileadmin/Publications/Tech_Papers/FXBlenderDoc_01.pdf (C:\0x\xz_ref)
outliers are particularly damaging to market making systems.
input – transaction prices and indicative quotes
output – only indicative quotes

— cross rate calculator

What we compute are backdated cross rates. PWM (non-trading) client agrees on a currency conversion AAA/BBB. She agrees on AAA amount, and wait for a few hours/days to get the actual BBB amount, just like cross currency credit card payment —

MasterCard exchange rates are based on multiple market sources (such as Bloomberg, Reuters, Central Banks, and others). These rates are collected during our daily rate setting process. The exchange rates displayed on the web site are derived from the buy and sell rates included in the MasterCard daily T057 Currency Conversion Rate File.

MasterCard applies the USD as unique reconciliation currency to manage all currency conversions globally. Due to possible rounding differences, the published calculated cross-rates may not precisely reflect the actual rate applied to the transaction amount when converting to the cardholder billing amount. The calculated cross-rates will be loaded onto the MasterCard web site on a daily basis.

MasterCard applies the exchange rate to transactions at the time of settlement, not at the point of authorization of the sale.

–ECN interface
(my code doesn’t use FIX. Our volume is lower. conversion from FIX to xml is encapsulated in a library)

FXAall, Currenex, Hotspot, TradeWeb , Bloomberg. Example applications:
* auto quoters
* price feeds
* market data feeds
* post trade feeds
* execution gateways

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s