A glance into algos and how they compare with the human elements of trading
Man vs Machine
upon a time in the financial world it was all about getting the right ‘man’ to
do the trading. Now, you may be forgiven for thinking that it is all about
getting the right machine.
impact of our new age has been most keenly felt in the two spheres of globalization
and automation and trading is no exception. The rewards can be large for those
able to capitalize on the benefits that automation offer to us.
example, the Medallion hedge fund founded in 1988 has
achieved an average annual return of around 30%. However, don’t get your hopes
up since the fund has been closed to the outside world since1993.
It simply focuses on trading its own money now.
In 2018 it had $84 billion assets under management with profits of around $25
What is algorithmic trading?
those unfamiliar with algorithmic trading it is simply trading that takes place
on an automated level. Computers are given specific instructions to follow
(algorithms) for making trades at large volumes and high speeds.
The largest portion of today’s algo-trading is
High Frequency trading (HFT) which places large numbers of orders and helps to
make liquid markets. The following are some of the most well-known algos.
Volume Weighted Average Price (VWAP)
executes a buy order in a stock close to its historical trading volume in an
attempt to reduce the trade’s impact on the market. To explain, imagine that
over a month 5% of a stock’s trading volume typically occurs in the first hour
with that knowledge then a computer with a client’s order will stop trading
that order as soon as the 5% level is reached. The remainder of the
order will be traded at a different time. The thinking behind this algo is to
disguise heavier than normal trading activity, so other traders/machines don’t
see what is happening.
If they did, and bid the price up, this would
impact the price at which the order was filled.
Trade Weighted Average Price (TWAP)
executes orders based on time. This is for the investor who wants to match the
levels of volume that are going on at any particular time. If there is an
increase in interest, then the algo will become more aggressive.
Similarly, if there is less volume going on then
the algo will become less aggressive. This is an algo used by momentum traders
who want to trade small, illiquid markets where volume analysis make’s less
Developed by Credit
Suisse it was developed to enter orders without signaling to the market place
that a large order is being placed. It has a variety of techniques designed to
cover its own tracks.
The algorithmic trader operating at pockets of volume
is the algo trader who find pockets
of volume in order to enter the market from the buy or sell side. This is type
of trading is most likely to occur in the stock market where volume flows can
type of trading is much harder to execute in the forex market especially for
retail traders. It is only those with
the ability to see large pools of volume that could profit from this knowledge.
This would be banks, large traders, and brokers with a good sight of the market
are supposed to be rules about front running these orders, but that is very hard to implement.So, the question you might be asking is, ‘Is
it possible to compete with algorithimic trading as outlined above?‘
short answer is no, not on an algos own terms. If the algo you are trading
against is a High Frequency Trader (HFT) scalping the markets with the aid of a
computerized program and advanced technology in order to aid execution, then
you can’t compete with that.
If speed is needed to enter after an economic
deviation then you can’t beat the algo for speed of execution. Furthermore, the
HFT will have considerable resources and will be able to keep the algo running
24 hours a day and 5 days a week. No-one can keep awake for that amount of time,
let alone function reliably.
Some algo trading uses technical areas to enter and exit
traders can still compete with algos by knowing when to take trades.
This is where an edge can lie for the old school trader. For example, some
algos will enter trades where pockets of large volume is likely to collect,
such as around the 100 and 200 MA.
When that happens, the man, can evaluate the
fundamental and sentiment of the market to allow the trade to run a little
further or even decide whether to enter or not. Take the GBPUSD currency pair
for example through December 13 to the time of writing on December 17. In the
GBPUSD chart below Theresa May had been struggling considerably in getting her
Brexit deal through Parliament.
a result there was no appetite to buy the GBP and all the rallies were sold.
Through December 13 and 14 Theresa May was trying to get assurance from
European officials that the Irish Border issue would not be allowed to drag on
indefinitely if the UK Parliament accepted May’s Brexit proposals.
was not prepared to give legal assurance to Theresa May and the bearish GBP
sentiment remained. In this instance the man has an advantage over the machine.
The man can enter orders with the knowledge that there are strong sentiment
factors to sell the GBP from the 100 and 200 moving averages on the 1hr chart.
The machine can only enter orders at the
technical level. The man can choose whether to enter to not and whether the
market dynamics are suitable.
There is still a future for the old school trader
So, the masters of
the old school still do have a future and it revolves around interpreting
market dynamics. There is also the human element that people like in the
finance world. A machine does not have a personality, whereas a trader does.
Some people will choose a man above a machine just out of preference for the
human factor that a computer can’t meet. Not yet, anyway.