Capitalmind
Capitalmind
Actionable insights on equities, fixed-income, macros and personal finance Start 14-Days Free Trial
Actionable investing insights Get Free Trial
Commentary

You Can’t Trick an Algorithm, says UK regulator

Michael Corsia, a High Frequency Fund trader, was fined $903,176 by the UK regulator for “manipulating” Oil futures by putting in large orders just to cancel them immediately.

The way he did it, using an algorithm (From Dealbreaker):

  • If the bid-ask spread is wide – say best buy is $95 and sell is $100 , he would put a small buy order at $96.
  • At the same time he’d put massive sell orders at $99, $98 and $97.
  • Other algos, thinking that this large sell volume will push the market down, would then rush in to sell at $96 – executing Corsia’s small sell order.
  • Immediately, Corsia’s algos would yank off (cancel) all the orders at $99, $98 and $97.
  • Now some other algo is left holding the bag at $96.
  • On the other side, Garcia’s algo would do a small sell order at $99, and then large buy orders in the 96-98 range.
  • Again, looking at the volume, other algos would eat up the $99 order, and lo and behold, the large buy orders would then get cancelled.

This is bad, to the other algos, the regulator has decided. However, it’s not just bad, it’s largely stupid, as Dealbreaker points out:

Why go after Panther and not all the other algorithmic traders cancelling all their orders? The answer appears to be that Panther, and its sole owner Michael Coscia, were completely nuts; as Paul Murphypoints out, Panther would bid for $100 million of oil in order to make $340 on its actual trades. It’s not entirely clear what Panther would have done if it’d been hit on its $100 million of fake bids but that sort of risk/reward ratio suggests the answer is “nothing good”; from a market stability perspective it’s probably for the best that they were canceled. Also from a market stability perspective it’s probably for the best that the CFTC and FCA shut this guy down.

Yes, the guy actually made just a profit of $279,920 (FT) from all that trading, and for his average $100 million dollar worth order, just $340. He was incredible lucky to not have Darwined himself out of existance – just one large move would have wiped him out. And he was, like FT said, half the oil market.

But this cancelling business happens all the time, even in India. If you have a terminal you can see the constant cancellations happening with prices being moved by algos. Retail traders are most affected (but largely these algos are also stupid, in that they can be gamed manually)

How would you prevent this?

  • For more than 500 cancellations per day, charge Rs. 1 per cancellation, increasing at every 100 step to reach Rs. 10 per cancellation beyond 1,000. (Replace Rs. 1 by 10 cents for the US). That’s the end of most HFT and algo stupidity right there. No retail trader will ever cancel that much.
  • Increase margin requirements on commodities. The tiny margins are reasons people do this kind of stuff in volume. Doing it in small quantities is irrelevant to anyone.

The current method, of fining small and inconsequential people, will do nothing.

Meanwhile, Goldman Sachs has cornered a large chunk of the aluminium market in the US and it seems, is moving the aluminium from one warehouse to another and holding it off from the market in order to earn more rental income.

Oh, you can quite happily manipulate aluminium buyers, because obviously Goldman is too big to fail.

  • ARP says:

    I have noticed a similar pattern in illiquid stock WABCOTVS couple of months back.
    If the spread is wide (say Rs 10 or more), and I placed an order inside the bid or ask, immediately an algorithm would place an order that was Rs 0.05 better than mine. Then (I guess) another algorithm would come in a place a order that was 0.05 better than the previous.
    The two orders would enter into a race and the price would improve by around Rs 1.50 (each algorithm effectively improving it own order by Rs 0.10 for half a dozen times).
    Haven’t seen it lately, but it was amusing.