June 21, 2017 at 7:51 am #24329
On the 2nd of June this year, exactly five days before the UK election, I wrote about Kensho, the market-driven artificial intelligence that supposedly has all the answers Google doesn’t.
Kensho had made a trade call regarding the UK elections, based not on the outcome, but rather on what would have made money over most of the last eight elections.
Here was Kensho’s prediction:
Buying the GBP five days before the election and selling it five days after would, on average, cost you 1.05%. So our trade was the opposite. We would test out Kensho by ‘selling’ the GBP on 2 June 2017, and buy it back on the 15th.
How did our Kensho paper trade go?
Here is the GBPUSD daily chart.
That 140 pip move from 1.2890 to 1.2750 is a 1.08% move. How about that.
Kensho’s other prediction was that a sell one day before, to be liquidated three days after would return 0.56%. Had you done that, you would have captured a 200 pip move….and even beaten Kensho’s prediction. Two for two.
The robot had one final call – suggesting that a buy on election day could be sold for a small profit a month later. We’ll keep an eye out for that one and see if the bot manages a hat trick.
Apart from the impressive results, there is another lesson here.
Kensho is a technical trader
Kensho is a “technical” trader. Kensho made its calls based on what happened over the last 8 UK elections. It didn’t pay attention to who was running, who was leading, or even who won. It ignored fundamentals and based its decision simply on what happened most of the time.
Kensho looked at the market and found a pattern.
This is exactly the job of the technical trader.
1. Identify reliable patterns in the market
2. Pull the trigger and take your trade
3. You can do exactly that
Rise of the machines
There is a second observation I would like to share. In the real world of trading, we have to consider more than just a direction. Direction (buy or sell) is a part of the puzzle. We also need timing (which Kensho admittedly provided), but finally, we need risk management.
Above: Kensho on a rare day off. (Terminator 3)
If you scroll back up and have a look at that GBPUSD chart I posted, you’ll notice that price went the “wrong” way for four days, climbing by nearly 100 pips.
Holding onto a short that has almost moved 100% in the “wrong” direction based off the expectations of a robot would have been tough. How far away would your stop have been? What if there was a genuinely surprising result in the election that sent the pound soaring? Would you have just held on indefinitely (and foolishly, in this trader’s opinion)?
When I first read about Kensho, I felt that it had value, and I still do. We’ll be following up here.
Practically, we made some money shorting the GBP over UK elections, also by following simple technical rules.
Could the robot simply be reminding us that what we already know “works”?
You must be logged in to reply to this topic.