One of many unanswered questions about the Sarao case is whether he might be able to reasonably argue that his alleged ‘spoof’ orders were actually part of a market-making strategy. Bradley Hope uploaded some more documents last night, that I think could help answer that question.
The FBI complaint mentions how often Sarao modified his orders, without much detail on the nature of those modifications. Modifying an order frequently doesn’t, on its own, suggest manipulative activity. There are lots of legitimate reasons why market-makers use that exchange instruction to update their quotes. And perhaps Sarao’s defense could make an argument that he was actually intending to market-make, only several ticks away from the BBO. He certainly would not be the only market-maker posting liquidity far away from the current price, a trading strategy that could be one of the oldest in existence. If so, it would be consistent with a high modification rate and low fill rate. And perhaps he could claim that he tended to post more quantity on the sell-side as part of his quoting strategy (remember, he appears to have mostly traded during volatile times).
So, one detail I would like to know is whether Sarao used those modification requests to update legitimate quotes as the market price moved, or whether he used them in some other fashion more indicative of manipulation. Last night’s documents may contain a hint. From p.53, which contains an email from Sarao to Matt Garley of Trading Technologies:
What i need are the following functions…
iv) The ability for my orders to rest on a particular size, ie my order will be pulled if there are not x amount of orders beneath it. Of course to make this work we will have to stay at the back of the book, v) this can be done by increasing/decreasing my order by a 1 lot every time a new order is detected where I am resting
This could help explain his allegedly high modification rate and, in my view, is really bizarre. If I understand it properly (and I might not), he is requesting a feature to automatically send 2 modification requests to the exchange whenever someone else joins his price level. The first request would be to increment his order’s quantity by 1 lot and, right after, a second request would decrement it by 1 lot. This would result in no net change in his order’s quantity, but would have the effect of moving him to the back of the FIFO queue. An order at the back of the queue only gets filled when the entire price level trades, usually indicating a very toxic incoming order.
Why would somebody want their order to be at the back of the queue, which almost certainly has worse adverse selection characteristics than the front? One possibility is that he wanted to be at the back of the queue in order for his strategy to have advance notice of the price level being swept-through, because trade confirmations sometimes are sent to those who were executed prior to being distributed on the public data feed. However, it wouldn’t make sense to send large orders for this purpose, nor is it consistent with the picture we have of Sarao being a low-speed trader. Another possibility is that he wanted to reduce his fill rate to avoid being filled simultaneously across multiple products, but still quote large size. I’m a little skeptical that doing so would improve a strategy’s profitability. But I’m not familiar with many markets, and maybe on some it makes sense for capital-constrained traders.
But there is a third possibility: that he wanted to be at the back of the queue in order to minimize his chances of trading, without regard for the actual profitability of trading. This would be consistent with the hypothesis that these alleged orders were intended solely to deceive others. Another hint that could support this hypothesis is contained in an email he sent to Joanna Jasina of the FCA (p48 of the same document):
No wonder they [HFTs] can manipulative on top of my orders without any risk, for even when I change my mind and decide to sell into my buy order, the manipulative orders on top of my initial buy order disappear in the 4 milliseconds it takes for my buy order to be cancelled and replaced with my sell order so that I do not trade with myself !!!!!!
This quote suggests that, when he pulled his orders, those behind him in the queue quickly pulled theirs. Hypothetically, if Sarao were indeed spoofing, this could indicate that he sometimes hoped to trade with those who copied his ‘spoof’ order. In order to not trade with himself, he would either have to delete his own order before trying to trade with the victims (evidently high-speed traders who could quickly delete in response), or he would have to be behind the victim’s order in the queue. I’m under the impression that the CME does not take kindly to self-trading and this was before the days of Self-Match Prevention, so, if Sarao was set on this type of spoofing, he’d have little choice but to send his orders to the back of the queue.
This discussion is all hypothetical, and I’m not sure to what extent the emails are actually reflective of what he was thinking, but I would like to hear more from investigators on the exact details of Sarao’s trading activity.