Correlation: when five positions are one trade

After this lesson you'll count your book's real number of bets, which is usually smaller than its number of positions.

The illusion of five

A trader runs five positions, each risking 1R, heat 5R. All five are semiconductor stocks. School IV, course 4 already told you what happens next: group members move together, because the same institutional theses drive the whole list. The day the group breaks, all five break together, usually on the same morning gap. The book holds five positions and one real bet, and none of its 5R of heat was ever diversified.

Correlation is the technical name for the degree to which two things move together. You don't need the statistics to manage it at this level; a rough count, done every time, does the job. Count your book's distinct bets. Positions in the same industry group are one bet. Positions in adjacent groups riding the same theme (the chip maker and the chip-equipment maker) are closer to one bet than two. Positions whose only connection is being growth stocks in the same market still share the regime bet, which nothing escapes.

The working rules that fall out:

• Cap the theme. However ranked the group, correlated risk gets a ceiling of its own — say, no more than 2R of heat in any one group or obvious theme. The checklist will keep pointing you at the strongest group's members, since that's where leaders live, so this cap tends to bind exactly when enthusiasm peaks, which is the point of having it.
• Name the bet before adding. Layer zero-and-a-half of the checklist: "what bet is this, and how much of it do I already own?" A fourth AI-theme position mostly presses the bet you already own. Pressing is course 3's subject and belongs there, with a plan attached.

The day everything correlates

The second lesson is harder and needs saying plainly: in a crash, diversification within stocks mostly evaporates. In the autumn of 2008, and again in March 2020, everything fell together, banks and railroads alike, because funds were meeting redemptions and leverage was unwinding, so holders sold whatever they could sell. The group a stock belonged to didn't matter that week. Correlations that measured 0.3 for years spend those weeks near 1.0.

So the theme caps are what manages ordinary times, and the extremes need something else: the regime response from School IV (exposure down as the gauges deteriorate), plus the overnight-event awareness from School II applied bookwide. The next course and course 5 build the drills. The summary of this course's two halves is that you count your real bets while the market is healthy, and you also size the whole book to survive the weeks when the market becomes a single bet.

Check yourself

1. Five 1R positions: three chip names, one chip-equipment name, one utility. Count the bets. (Roughly two and a half: the four semiconductor-theme names are one heavily-pressed bet (~4R), the utility another, and all five share the regime bet.)
2. Why will the theme cap bind precisely when you least want it to? (Leaders cluster in the strongest groups, so enthusiasm concentrates there — the cap exists to interrupt that exact concentration.)
3. What did 2008 and 2020 teach about stock-picking diversification? (In liquidity crises it thins toward nothing — everything falls together when the selling is about money itself. The defense that still scales in those weeks is the regime response.)

The habit this lesson installs

Before every entry, name the bet and count how much of it you already own.

Next: Course 3 — "Progressive exposure."