minimax

It doesn't mean anything for Chicago. "NYSE Chicago" is an electronic exchange hosted in New Jersey. It has Chicago in the name because NYSE bought the old Chicago Stock Exchange in 2018 (which was partially hosted in Chicago and partially in NJ... it was weird). To my knowledge post-2018 NYSE Chicago never had any trading operations in Chicago. NYSE does have sales people in Chicago because of the large prop trading community in the city and it would be hard to see how this NYSE Texas situation changes that.

Your description of market making is backwards. If you're buying the ask and selling the bid, you're paying the spread not collecting it.

It doesn't happen. All the exchanges have a "Day" order type that you can send before 9:30 that will be live on the book when it opens at 9:30 (or transitions to the "core" session at 9:30, most US exchanges have a premarket session prior to that). The idea of having some sophisticated strategy that sends 100ns before 9:30 is nonsense.

The NYSE runs out of a public data centre (called NY4) which is run by Equinix.

No. NY4 is in Secaucus. NYSE operates out of an ICE (NYSE parent co) owned facility in Mahwah about 25 miles north of there. They managed to pick out the one big US equities exchange operator _not_ running in an equinix facility.

Sorry but this whole post sounds like someone who is sort of HFT adjacent but doesn't really know what they are talking about. Sending orders at "09:29:59.9999971 at the hope your order arrives at 100ns past 9.30am." What?

if this was found by accident, how many things still remain undiscovered.

This, to me, is the most important question. There is no way Andres Freund just happened to find the _only_ backdoored popular open source project out there. There must be like a dozen of these things in the wild?

Yes, exactly. Fidelity is a private company just like Reddit and Discord are. There are still shares in the private company held by various owners, but the shares aren't traded publicly, which is what it means to be a public company.

Fidelity is not a public company.

What lead you to investigate PCIe relaxed ordering? Can you suggest a book or other resource to learn more about PCIe performance?

Why did you feel the need to support trading of non-LTSE listed names on the LTSE? Are there novel order types or other trading mechanics coming down the pipe?

You can't really "front run" a retail order. Front running means you (assuming you are an agency broker) get a big market moving order from a customer, do a trade in your own account first (in the same direction as the customer order), then execute the customer order which moves the market in your favor.

Retail orders are small and generally aren't capable of moving prices. Your order for 100 shares of whatever isn't going to move the price so you can't really make money ahead of it.

Can anyone suggest a good book that would serve as an introduction to finite field arithmetic? I keep randomly running into it (e.g. this post), but don't understand it well enough to follow the discussion.

Today is Feb 21, 60 days ago is late December not November.

The S&P 500 is up 15% from 60 days ago which is considerably more than all of the positions in your screenshot.

All the transactions detailed in the SEC order are outright sales of stock (not options). He was both selling stock he already owned to avoid losses and selling short to generate additional profit. Before it was acquired, AFOP never had a very substantial market cap so it’s possible AFOP didn’t actually have listed options, or that it did but they weren’t liquid.

I suspect that AFOP was probably a typical low volume small cap and that Li’s selling while not large in nominal terms was probably still pretty significant relative to the average daily traded volume. Large enough to be picked up by a pretty simple screen anyway.

The XIV is (was) meant to track the daily inverse of (roughly) the same underlying index as VXX. So a 10% up day for the index should be 10% up for VXX and 10% down for XIV. Now we can plug in rough approximations for Monday 2/5 and Tuesday 2/6.

The underlying index was up something like 90%. So VXX up 90% and XIV down 90%. Next day the index goes down 25% so VXX down 25% XIV up 25%. The two day returns for VXX will be 1.90 * .75 = 42% up and the two day returns for XIV will be 0.10 * 1.25 = 87.5% down. You can see how the daily tracking blows out the tracking over longer terms (just 2 days in this case). Since the VIX moves had been relatively small on a day to day basis, it sort of looked like they tracked each other on inverse terms over longer periods but it was just an illusion.

Lots of people trade physically settled commodities futures contracts without handling the underlying products. You just have to make sure you close or roll your position before it settles.

The article isn't about trade throughs. It's an unconvincing argument that rebates bad and that IEX is good because it does not use rebates (while failing to mention any other exchanges that don't pay rebates).

How is that a fix? A market with huge displayed sizes but no trading (because nobody wants to pay the massive take penalty) is not a "liquid" market.

0.5% on equities was the Bernie proposal. Here I'm assuming that both sides pay 0.5% but even if it's half that (each side paying 0.25% or only one side pays) the numbers are crazy.

Here are some numbers. 0.5% of a $50 stock is $0.25. So to break even on the tax alone you need to sell $.50 higher than you buy. That alone will blow out the spread any market maker is able to quote at. The other problem is that now scratching (you buy at the bid and now it looks like the price is going the other way so you aggress and sell back into the bid for no profit) is also extremely expensive (you lose $.50 per share on a $50 stock just scratching). That's going to really kill your profitability. Maybe someone could figure out how to make it work, but it would be an extremely painful regime for market makers.

The real reason the article is unserious is because IEX is basically purpose built to game the execution quality metric they're looking at (effective spread... trade price vs NBBO midpoint). Imagine you built a dark pool that only executes at the midpoint of the NBBO. It would have a perfect "effective spread" score but contribute nothing to the displayed NBBO. That's essentially what IEX is. Their midpoint order type only has value because of the displayed liquidity at the other rebate-paying exchanges.

I think that's about half right. A 0.5% tax on all equities trades, which is what Bernie was proposing, would put a bunch of HFTs out of business basically overnight and at the same time dramatically increase costs for investors via 1) the tax (obviously) 2) wider spreads 3) reduced liquidity.

If my Robinhood trades of 100 lots of more execute at a few pennies higher than the price when I place the order

They're not really allowed to do that and I don't think they are doing that. Robinhood (or whatever wholesaler they route your order too) is required to execute your order within the nbbo. They can't just e.g. fill your buy orders a few cents above the national best offer because they want to make some extra money. The regulations don't work that way...

Any reasonably good optimizing compiler will implement that switch statement as a jump table so it's really just one conditional branch even though it doesn't look like it.

My understanding is that nothing serious has gone wrong, so far, in nearly two years.

GBTC is a bad way to trade bitcoin and the chart in reference [2] you posted shows exactly why. A good BTC tracking ETF would have NAV = market cap, but here you can see that NAV and market cap are unlinked (as are returns for GBTC vs bitcoin) and people are generally paying way over NAV (in some cases over 50%!) to own GBTC. This is presumably due to unsophisticated traders buying GBTC without understanding that they're getting a terrible deal.

In summary, yes something is seriously wrong with GBTC and has been wrong for the entire life of the product.

What calling convention passes the first integer argument in rax and the return value in rdi? Thats exactly backwards from what I usually see.

There is no non-voting FB stock yet. It's still tied up in a shareholder lawsuit.

Ah I see. In size() and c_str(), strictly speaking it's not the category constants that become 8 byte immediates but rather the category mask. Since the category constants and masks are all compile time constants I wonder if a smarter compiler could have performed this optimization for you.

That is a smart optimization but it does make you wonder why were they using a size_t to hold the category (only 3 possible values) in the first place...

Are any of them doing multicast market data yet?