HFT on TOM : whodunnit?
Half a year ago I reported on the clash between DeGiro and Binck, with respect to the not-so-smart order router from TOM. See DeGiro trashes TOM and Flash Boy Binck ignores law of holes. Few months later TOM decided to shut down their stock trading venue. Assumed that would be the end of the story.
However, great news! Radio show Argos (nl) got interested in the matter. Somebody read a book from Michael Lewis, and without prior knowledge about the stock market they started investigating. Had a coffee with both of them, and to be honest the result looks good. They managed to get a story about HFT trading even my mum will understand. See below for the radio show with “visuals” on youtube. It’s still only in Dutch, understood the English subtitles will arrive next friday. (update : arrived, see below)
So, whodunnit?
The stock market is a strictly anonymous game. Still I’m pretty certain to know which HFT shop managed to jump ahead of Binck’s retail flow. First of all, you need to have raw speed to be able to do this kind of arbitrage. Second, my friends at IMC and Optiver confirmed me they weren’t involved in this. Someone else also suggests both of them are lacking experience and manpower to do this game. It’s mainly an American game. That reduces the number of suspects.
Marginal business
There a many US high frequency trading shops active in Europe. Tower, LaTour, Spire, Hudson River, Citadel to name but a few. Only one is active in Europe with a real significant trading team with a group profit and loss. Other firms with smaller outfits in Europe aren’t active in this kind of marginal business. We’re not talking about a cash cow here. Also looking at TOM’s member list at the time, one suspect remains. Though the member list isn’t saying everything, sponsored market acces is possible.
Everything points at Virtu
Everything points at Virtu. Yes, the very same firm who did an IPO in the meantime, after postponing it earlier due Michael Lewis’ book “Flash Boys”. The controverse coming from the book about the HFT firms blocked the first IPO attempt. However, Virtu would never comment on this and nobody will be able prove it. In the meantime reliable sources are confident it was Virtu.
Latency arbitrage is fine
No matter what politicians said on the radio, latency arbitrage by firms like Virtu is perfectly legal. Selling on TOM and rapidly scalping on Euronext is fine. Or other way around. Every trader would do exactly the same. If you can sell a stock on venue A and buy it back at a lower price on venue B ; money in the pocket. Just the speed makes it different to grasp.
The problem has always been TOM’s order router which couldn’t handle large orders. And they made a promise they couldn’t keep. There’s nothing in it for them to do it on purpose.
Back to Argos
The journalists, Hansje van de Beek and Stefan Heijdendael, managed to have some high profile US experts look at the discussion. Eric Hunsader and Brad Katsuyama himself checked the trading tickets and confirmed it looked exactly like Michael Lewis’s book Flash Boys. Nothing new here, except for the fact they managed to get other people react.
TOM stock venue is gone
Binck responded in a classical silly way (nl – something like ‘we are not HFT, DeGiro is because HiQ is HFT!“). Anyway, the TOM stock market is closed so everything is a little bit an academic discussion. A thing of the past. The TOM option market is still alive and kicking, and as far as I can see it works as precise as a Swiss clockwork. No HFT is taking advantage of option orders on TOM.
first, here he we go with sunday evening post, what’s with jack and his routine
Most people are busy with work during the week and so Jack puts stuff up when he has time on the weekend.
The two most miserable commenters on AT are the “complain posts go up on Sundays” guy and the “who cares, next” guy.
good one, jack. keep them coming.
It would have been very easy for HiQ to build an algorithm that would do exactly the flash trades in their TOM experiment, given that they knew that would be placing relatively large orders in specific illiquid stocks. Especially since you could reasonably expect that not too many other people (if anyone else at all) could interfere with your silly orders. In other words it cannot be ruled out that DeGiro faked their TOM experiment with help from their own algo desk at HiQ.
Possible, but unlikely.
1. Trades were done in presence of a FD journalist.
2. HiQ hedge fund is absolutely not a fast HFT.
3. TOM SOR being not top of the bill – who is surprised?
4. AFM would know.
5. Even if they are able to fool everyone (which i think they can’t), DeGiro is not evil.
So, probably virtu just walks away with the not illegally earned money. How much was it, anyway?
1. And what does that prove? Nothing. Or are you trying to say that out of the entire universe of stocks, the journalist was allowed to choose the stocks and quantities that DeGiro would send to TOM?
2. You don’t need to be the fastest. You only need to program your algorithm that as soon as you are filled on your first bait order (= DeGiro has placed its order), you quickly clear out the order books elsewhere, so it looks as if a “flash boy” is active. Anyone with co-location can accomplish this faster than a broker will who only react on subsequent order book updates.
4. The AFM would not be allowed to comment / make that knowledge public.
Can anyone confirm/deny that Tibra made 100 mln + this year?
They must be on a massive profit streak if this is true.
when they mail you your dividend cheque, you’ll know for sure, till then, stop wasting time on rumor mill
> Everything points at Virtu.
Sorry, Jack, I only got to read this now. I don’t get it. Which “everything” are you talking about? Please give some examples. Or every example. Or am I missing out on something?
“Which “everything” are you talking about? ”
Re-read the paragraph under the heading ‘Marginal Business’
Ah, nice use of http://tldrfy.com . I love that little Chrome extension
tldrify.com – https://chrome.google.com/webstore/detail/tldrify/dbphpdgmhigmaepjklmklmlcoinihjdo?utm_source=chrome-app-launcher-info-dialog
this whole blog is about ‘being pedantic’, glad you finally figured it out
A lot of people don’t realize how fleeting a game many strategies really are. At one desk I was on, we had one team that had built a pretty effective F/X latency arb strategy. Once they got it up and running, they were clearing $50k a day by essentially trading against “stale” quotes.
Of course, it was only a matter of time until the brokers recognized what was going on, and moved to tighten that all up. Within a few months, PnL was down to $100 a day at best. This didn’t stop the guys from clearing a few million dollars, but it wasn’t a license to print money forever.
how did you conclude that other people dont realise how fleeting a game this really is?
“The two most miserable commenters on AT are the “complain posts go up on Sundays” guy and the “who cares, next” guy.”
I’d like to add the “how did you conclude” guy and the “define” guy.
Although they’re all probably the same person.
To save him the trouble of doing so, I’ll ask the question myself: how did you conclude that they’re the same person?
Ah well, this blog isn’t what it used to be. For sure someone at Tibra can post the ytd Numbers. And just as interesting: how are others doing? Liquid? Susquehanna? Virtu? Citadel? All options? Eclipse? Ark? Optiver? Imc? Maybe even scrocca? 😉
Earth to Bhandari: Virtu (“VIRT”) went public in April and published their Q1 results two weeks ago. They’re doing very, very well.
how did you conclude its a him?
To save you the trouble of doing so, I’ll like to add that’s it’s maybe you: a proper wanker and that’s why?
why do you care on how much tibra or optiver or any other company is making this year
shouldn’t you just look at your own pnl and be happy with your miserable life rather than see somebody else’s faster car and be even more miserable
Does anyone know how Scrocca is doing ? I heard a new group is leaving the company…Probably not because company goes well…