Thomson Reuters Adds 15 Pairs to Faster Data Feed

Thomson
Reuters has added 15 new currency pairs to its binary multicast feed for
Matching.

Profit & Loss previously reported
in October 2016 that the firm was planning to increase update frequencies for
Matching real-time market data by up to 10 times to 25 milliseconds.

In phase
one of the roll-out Thomson Reuters put USD/CAD, USD/SGD and EUR/USD pairs in
the binary feed on May 21, 2017. Then on Sunday it launched phase two, adding
the following 15 currency pairs: AUD/USD, AUD/NZD, NZD/USD, GBP/USD, EUR/GBP,
EUR/DKK, EUR/NOK, EUR/SEK, NOK/SEK, USD/MXN, USD/CNH, USD/HKD, USD/JPY, USD/CHF
and USD/RUB.

“The client
feedback from the initial launch has been very positive and there’s been a very
big demand to get these additional currencies on,” Paul Clarke, head of FX
trading venues, tells Profit & Loss.
“The intention is that as clients get more updates in terms of where the book
is, we’ll start to see activity coming in more frequently, and we’ve actually
already started to see this.”

In
parallel, TR is also adjusting the minimum quote life (MQL) for the 15 pairs
added to the binary feed down to 50 milliseconds, with Clarke explaining that
there’s always a balance between market data frequency and how long clients are
expected to leave orders in a platform.

“Given the
feedback we’ve had from our clients, our own data analysis, the rules that we
have on the platform and the fact that we introduced a randomisation mechanism
last year, we don’t think that speed is the only thing that impacts clients’
trading behaviour. The MQL is less relevant now in terms of getting a really
good trading experience,” he says, highlighting TR’s randomisation mechanism as
something that has provided the firm with the flexibility to speed up market
data on Matching.

Thomson
Reuters is also planning to speed up the market data for another set of
currencies on the standard Matching API in September. Currently, the most
liquid instruments on the Matching API feed update every 100 milliseconds,
while the remainder update every 250 milliseconds. The plan is to bring a set
of additional instruments up to the same speed as the most liquid ones.

Over time,
Clarke says that TR plans to analyse the data from the binary feed to judge its
impact, and then it will decide whether to add more currency pairs onto the
feed or change the frequency of the feed.

Asked what
he expects the data to show, he explains: “If we look at activity on the
platform, we typically get peaks of activity after the market data has been
updated, and so obviously one behaviour we would expect upon increasing the
speed of the market data update is to see more of those peaks through the
second.

“The other
thing that we’ll be analysing is what we see in terms of the book and in terms
of the spread,” he continues. “Because we’re slightly reducing the minimum
quote life, we’re anticipating there might be a reduction of spreads throughout
the day.”

Clarke also
explains that the company, as a matter of course, constantly manages the
granularity of its instruments and makes changes where it feels necessary. For
example, over the past 18 months it has changed the granularity of the
Scandinavian currencies, going from one pip to five pips.

“Granularity
is one of the key levers to make sure that we’ve got the right client
experience,” notes Clarke.

Commenting
on the factors that determine whether TR will change the granularity on a
currency, he says: “We look at what is the value of a pip or a basis point. So
effectively, what is the value of a single tick in terms of what is the minimum
that a client can change the rate by and what’s the value of that to $1 million?
And then, what’s the ratio of that to spread? We also look at certain
behaviours, so in terms of spread we look at percentiles through different time
zones of what the spread typically is for that currency.”

One
characteristic of the modern FX market that has been brought about by greater
granularity is a proliferation of typically smaller high frequency market
making firms that repeatedly place a bid of offer just inside the current top
of book. This can cause frustration amongst platform users, as the orders are
typically small in size and exacerbate signalling risk.

“When
deciding whether or not we should have less granularity, we do look behaviour
around the top of book,” says Clarke. “For example, how often is there another
offer that immediately comes inside it, what we call tagging?  We also look at how often we see that
behaviour occurs compared to joining – where another participant bids or offers
at the same level? We look at those statistics in terms of behaviour, because
obviously that has a big impact on user experience, if you’re leaving a resting
order and someone immediately comes inside, that’s not a great experience.”

Colin Lambert

Share This

Share on facebook
Facebook
Share on google
Google+
Share on twitter
Twitter
Share on linkedin
LinkedIn
Share on reddit
Reddit

Related Posts in

Profit & Loss is no longer publishing

Thank you for 21 great years of support