Who Else Is Here

Everyone in the book is there for one of four reasons, and the reason predicts the behaviour far better than the size does. Sorting people that way also tells you where the money goes — including whether any of it is coming to you. On one day of the instrument lesson 2 printed, the four and a half million uninformed shares hand over $22,500, and it arrives as $5,000 for the quoters and $17,500 for the informed. Not approximately.

Prerequisites: Lesson 5, for the two risks a quoter carries, and lesson 8, for why a signed print tells you about impatience and not about identity.

One day, in dollars

Take the instrument lesson 2 printed, quoted 50.02 bid and 50.03 ask, and a day in which five million shares crossed that spread — five million shares where somebody was unwilling to wait. A penny wide means crossing costs half a cent a share against the midpoint. Suppose a tenth of that crossing volume was informed, and that on those trades price went four cents the informed trader’s way before the quoter could get out.

Everything below is those four numbers and nothing else.

Who	Paid	Received	Net
Uninformed (4,500,000 shares)	$22,500 in spread	—	−$22,500
Informed (500,000 shares)	$2,500 in spread	$20,000 from the move	+$17,500
Market makers	$20,000 to the informed	$25,000 in spread	+$5,000

The three lines sum to zero, which they have to. But the useful reading is the vertical one: the $22,500 the uninformed paid is exactly the $5,000 the market makers kept plus the $17,500 the informed took. Not approximately — exactly, because there is nowhere else for it to have come from.

Notice what that does to a familiar complaint. The market makers did not take the informed traders’ money and they did not lose to them either; they collected twenty-five thousand and handed twenty back, which is the business lesson 5 described. The transfer that actually happened was from the four and a half million uninformed shares to the five hundred thousand informed ones, and the quoting firms were the plumbing.

Every figure in that table came off this module. The half-cent is the penny spread lesson 2 printed, halved because you cross to one side of it. The four cents is the move lesson 5’s quoter is pricing against, and the reason they cannot simply widen until it stops hurting is lesson 4’s arithmetic: the spread that makes them safe is the spread that makes you unable to trade. The five million shares were sorted into the first two rows by lesson 7’s rule, which is wrong about one in five of them, so the split between those rows carries lesson 8’s interval around it. And a candle of that day would have shown you five numbers, not one of which is anywhere in the table.

So the question the whole module has been building towards is not “what are institutions doing”. It is: on this trade, which of the four am I? If you cannot say what you know that the person on the other side does not, the accounting above has already assigned you a row.

What this does not settle

That you can tell which is which from the tape. You cannot, and lesson 7 and lesson 8 are why: a print carries size, price and an inferred sign, and a hedge, a rebalance and a conviction bet all look identical in it. The taxonomy predicts behaviour in aggregate and over time — when hedgers trade, how quoters respond to inventory — not the identity behind any single fill.

Nor are the figures above a measurement of your market. They are one internally consistent set of numbers chosen to make the structure visible; the informed fraction and the size of the move are exactly the two quantities that differ between instruments, and problem 3 is a way to bound them for yours rather than assume them.

And “informed” here means something narrow: holding a view the current price does not reflect, and being right about it often enough to pay for the times you are not. It does not mean insider, and it does not mean sophisticated. A great deal of confident, well-researched, expensively-produced trading is noise in this sense, and the rest of this course is largely an argument about how you would ever know which yours is.

The table is also not the whole outflow. Money leaves the four rows by routes that appear nowhere in it: commissions, exchange and clearing fees, financing on anything held overnight, and tax. None of those is a transfer between the rows; all of them are a transfer out. So the $22,500 the uninformed paid is the floor of what that day cost them, and the whole of module 2 is the arithmetic of the part this page left out.

And four is a useful number, not a measured one. Nobody counted a real market and found four kinds of person in it. The division earns its place because each role implies behaviour you can look for and the four together close the accounting, which is a good reason to use it and not a reason to believe it is complete. A fifth role would close the accounting too, and would explain things these four do not. Nothing on this page rules one out.

Four reasons to be in a book, and one of them is paid for by the others. Module 1 has shown you the machine; the question it leaves you with is which part of it you are.

Problems

1. Move one number. Rerun the day with the informed fraction at a fifth instead of a tenth, holding the five million shares, the half-cent and the four cents fixed. Give all three net figures, check that they still sum to zero, and say at what informed fraction the market makers stop making money at that spread.
2. Name the role. For each of these, say which of the four it is and what in the description settles it: a fund selling on the last trading day of the month; a firm that quotes both sides of forty instruments and goes home flat; someone buying because the chart broke a level; someone buying because they have read the filing and think tomorrow’s number is wrong. Then say which one you were on your most recent trade.
3. Bound your own. On your instrument, take a bar with a large signed delta and measure how far price travelled in that direction over the following ten minutes. Do it for twenty such bars. The typical move is your version of the four cents; a median near zero means the crossing flow you are watching is mostly not informed, whatever it looked like at the time.

Sources. Walter Bagehot, “The Only Game in Town” (Financial Analysts Journal 27, 1971) — the pen name Jack Treynor published under — the essay that first set the dealer between informed and uninformed traders and wrote down who pays whom. Fischer Black, “Noise” (Journal of Finance 41, 1986), on why trading that is uninformative is nevertheless what makes a market possible. Larry Harris, Trading and Exchanges (Oxford, 2003), on why people trade, which is the long version of the table above.

That is the mechanism, end to end: a price, a book, a fill, a spread, the person who sets it, what your screen does with all of it, and who else is in the room. The next module asks what the whole thing costs, and starts from a fact the mechanism has made unavoidable: every position you will ever open starts at a loss, with one of four charges deepening it for as long as you hold.

Educational only. Trading involves substantial risk of loss. Not financial advice. Past performance does not guarantee future results.