$10,000 in Micron since 2014. The same stock, played two ways that never overlap for a single minute. One loses more than half the capital. The other multiplies it almost 100 times. The difference isn’t which stock to pick — it’s whether the market was open or closed while you held it.
The experiment: two ways to play the same stock
Take Micron since 2014 and split every trading day into two segments that never touch: the session (open to close) and the night (close to the next open). Then play each segment separately, every day, for twelve years.
$10,000 invested since 2014, no commissions
Market hours only (buy at open, sell at close) — $4,512
Off-hours only (buy at close, sell at the next open) — $978,163
Playing only the session, you lose more than half the money. Playing only the night, you multiply it almost 100 times. All the return Micron has produced in twelve years happened while the market was closed — during market hours, the stock lost money.
Session and night aren’t symmetric, nor independent of the previous day’s close — they’re two series that, multiplied together, reconstruct the stock’s total move. One of the two concentrating almost all the return isn’t impossible: it’s just an unusual way of splitting up the same final result.
Now with real commissions
This experiment means buying and selling every single day — 6,360 trades over twelve years. The lines in the chart above carry no commission. With ProRealTime‘s real commissions ($0.005 per share, $1 minimum per trade), the picture changes quite a bit.
The same $10,000, with real commissions applied
Overnight strategy (buy at close) — $438,723
Doing absolutely nothing (buy and hold) — $441,309
Intraday strategy (buy at open) — runs out of money on January 21, 2021
Without commissions, the overnight strategy multiplied the capital almost 100 times. With 6,360 real trades in the mix, it ends up practically tied — even slightly below — simply not trading at all. The intraday strategy, meanwhile, doesn’t just lose: it goes completely broke before reaching the end of the period.
What this teaches about the cost of trading a lot
$0.005 per share with a $1 minimum sounds like nothing. But 6,360 trades accumulated over twelve years eat up, in this specific case, all the edge that timing seemed to offer. It’s not that the commission is high — it’s that the number of times you pay it is enormous, and the effect compounds just like returns do.
This is Micron, not a law of the market
This split between session and night is specific to this stock and this particular period — it isn’t a pattern you can extrapolate to any asset, nor a recommendation to “only trade at night.” Every stock, in every time window, splits its return between session and night in a different, largely unpredictable way.
Past performance doesn’t guarantee future returns. This exercise compares a real, verifiable historical case, but it isn’t a suggestion about what to buy or how to trade. The goal is purely educational: to show how a stock’s return splits between session and night, and how real commissions can erase an edge that looks huge on paper.
Conclusion
In Micron, since 2014, all the return has arrived while the market was closed — during session hours, the stock lost money. But that nighttime edge, so clear on a commission-free chart, dissolves almost completely once you trade with real costs: it ends up practically tied with doing nothing. The lesson isn’t “buy at the close and sell at the open” — it’s that any timing edge has to be measured after commissions, not before, because those are two numbers that can tell completely different stories.
