Read the Open
Posted (ET)
Access
Free
Filed under
Course, Index dip

The Close, lesson 2: Why the money is made overnight

Where the profit in a dip trade arrives, why the rule holds through the night, and what that costs.

Most of a dip trade’s profit arrives while the market is closed. That is where the rule is exposed, and it is why the rule sells at an open.

Day Overnight 09:30 16:00 09:30 Rule buys, just before the close Rule exits Every trade is held through at least one night
A. One day, to scale. The cash session is the short block on the left. The hatched night is the rest of the clock. The rule buys just before the close and exits at an open.
Rule buys Rule exits at the open Night Night Night Day Day Day Day Illustrative, not data
B. Illustrative, not data. Price drifts in the trading days and most of the rise arrives in the nights. The rule buys at the close of the first day and exits at the open of the fourth.

Rules

  1. Split a trade into its nights, close to next open, and its days, open to close. Judge each part on its own.
  2. The rule holds through the night on purpose. It buys just before the close and sells at an open, never into a close.
  3. When the rule says sell, it holds that last night too. The sale fills at the next open, which is Monday’s if the signal came on a Friday.
  4. The night is risk as well as return. The rule has no stop, so a gap against it is taken in full. Lesson 6 covers size.

What the evidence says

Across 25 to 33 years of index ETFs, the nights in a dip trade earned +0.38% a trade and the trading hours +0.12% (t 7.3 and 1.7): about three times as much overnight. Since 2010 the trading hours earned +0.06% (t 0.6), which is about nothing. On MNQ from 2022 to 2026 the nights made all of the profit.

So the rule holds the last night. Selling at the next open instead of at the close after the signal raised the average from +0.455% to +0.520% a trade (t 5.60 to 6.17), with the same trades and no extra trading cost. It did better than the old exit on three of the four ETFs, measured by return per unit of drawdown, in both halves of the history. A second, independent rebuild of the test got the same numbers.

Then we checked it on bars that no test had touched: 17 trades on intraday Nasdaq futures data from 2020 to 2022. Per NQ contract, the new exit averaged about $1,600 a trade against about $900 for the old one. That is a small sample, so it agrees with the ETF result rather than proving it.

It has a cost: one more night of exposure per trade. On intraday bars the worst trade grew by about a third. Holding only the nights did worse than the full trade (about +0.3% a trade), so the rule needs both parts.

The tests show where the profit arrived. They do not show why, and a rule should not depend on a story.

Hypothetical results1, not actual trading. The full statement is at the end of this page.

Exercise

Take the five dips from lesson 1. For each, split the move into its nights, close to next open, and its days, open to close, and add each part up. Which part carried the trade? Write down what you expect before you look.

The pointMost of a dip trade’s profit arrives overnight, so the rule buys just before the close and sells at an open, never into a close.

Check yourself

  1. It is Friday, you are in a dip trade, and the rule says sell. When does the sale fill?

    1. At Friday’s close, to avoid the weekend
    2. At Monday’s open
    3. At the first price you can get on Friday evening
    Answer

    B. When the rule says sell, it holds that last night too. The sale fills at the next open, which is Monday’s if the signal came on a Friday.

  2. To avoid overnight risk, a trader holds dip trades only during trading hours and is flat every night. What does the lesson say about that?

    1. It keeps most of the profit, because the trading hours carry the move
    2. It makes no difference, because the two parts earn about the same
    3. It gives up the part of the trade where most of the profit arrived
    Answer

    C. In the tests the nights earned about three times as much as the trading hours, and since 2010 the trading hours earned about nothing.

  3. Selling at the next open instead of at the close after the signal raised the average trade. What did that change cost?

    1. One more night of exposure per trade, and a larger worst trade
    2. Extra trading costs on every trade
    3. Fewer trades in the test
    Answer

    A. It used the same trades with no extra trading cost, but it added one more night of exposure per trade, and the worst trade on MNQ grew.

Next: The exact rules

Education only. Nothing here is advice about your account.