Support in a downtrend is a level to watch, not a setup

There is a trade that looks right every single time it appears, and for more than two years we kept taking it.

A good business. A long decline. A price level on the chart that had held before. A few rising histogram bars suggesting the fall had finished. And a premium rich enough that being paid that well to take the other side felt like the whole point of selling puts.

That was the method. Sell puts at support on quality names that had recently fallen a long way, with early strength as the entry signal: a handful of rising bars after the decline, at a level that looked solid.

Too often, those names were in heavy downtrends that simply continued. Too many kept falling, some of them heavily, and positions went deep into the money. This was not one trade or two. The experience ran across many names.

What follows is what was wrong with it. Two mistakes, neither of them the strike, and a third thing that happened afterward and made both of them worse.

The first mistake: a recommendation treated as a decision

Several of those names came to us as recommendations from a paid research service, which we do not name. They looked like quality businesses.

We took them because the service was paid for. We never ran them through our own screen.

That is the whole failure, and it is awkward in proportion to how ordinary it is. A subscription creates an obligation in the mind of the person holding it. You paid for ideas, so an idea arriving feels like a return on the payment, and checking it feels like an insult to what you bought.

The rule that came out of it is one line. A recommendation, paid for or not, is a candidate and not a decision. It goes through the same checklist and the same due diligence as a name you found yourself. The screen we did not run is the screen that would have stopped those trades at the first item.

The second mistake: early strength is not a turn

The charts already carried warnings when those trades were opened. We read them, and we found what we had come to find.

A few rising histogram bars after a long decline are early strength. They are not a trend that has turned. The first higher low had not formed, and until that sequence completes and the trend reads turn with it, the honest answer is not yet.

What the trend test is actually asking

Not whether the fall has paused. Whether a higher low has formed behind the pause.

Illustrative. Both panels are drawn by hand to show the shape of the test. Neither is a real security's price history, and neither is a recommendation to trade anything.

Support in a downtrend is a level to watch. It is not a setup.

Every piece of the picture invites the trade, and the trend test fails it. That is exactly why the trend test has to come before the level rather than after it. Read in the other order, the level arrives first and the trend read becomes something you argue with, and a trend read you are arguing with is one you have already decided to overrule.

When the richest premium argues with the trend read, the trend read wins. That sentence is easy to agree with and hard to follow, because the premium is a number on the screen and the trend read is a judgment.

And then the rolls went where the trouble was

One more thing those charts showed, and it is the part that surprised us.

Both opening trades stayed clear of earnings, as the rule requires. Both were rolled once they were already in the money, and the rolls carried them into the earnings window we had been careful to avoid.

There is nothing unusual in that. A trade that goes against you gets rolled into whatever sits ahead of it, earnings included, because by then you are choosing between the dates that are left rather than the dates you wanted.

Which means the margin of safety that the fundamental screen gives you is bought before the trade, when you still choose the dates. It is not available later, at the moment you would most like to have it.

The order that replaced it

We changed what we select. Trending, high quality names now, even at lower premium.

And the checks run in one fixed order, never the other way around.

1. The tide. The broad market first, before any candidate is looked at.
2. The trend, on the candidate itself.
3. The levels.
4. The steadied sequence. The bar that completes it is the entry.
5. The trigger. The one step in the path that is a nice to have. Kept in its place, because when it fires it sharpens the entry the steadied bar already gave you.
6. The strike position, last.

Inside that path the oscillators give one read, and the read has a name: resumption. The fall has stabilized, and the trend is resuming. The minimum version is the MACD histogram, the standard 12, 26, 9, ticking up from the pullback low with price back above the minus one ATR line. The preferred confirmation is the 14 period Relative Strength Index recovering from under 30, or turning up from its own pullback low. When both speak, confidence rises without adding another tool.

Two regime rules sit above all of it. Trend following tools are ignored in a range, where every crossing is a false signal. Oscillators are ignored when a trend starts, because they will spend the whole trend overbought. And an overbought reading matters here in exactly one way: as a reason to wait.

The chart is rarely perfect, and the path does not ask it to be. A trader who waits for every line to agree trades rarely, and that is a cost rather than a virtue.

The part that is uncomfortable to write

We departed from those selections and changed the method. And the honest part, said plainly: a trader with less capital, running the same method over the same two years, could have been devastated by it.

That is not a caveat at the bottom of a page. It is why the rule appears in block after block of what we teach, and why this article exists instead of a cleaner one about the trades that worked.

If one thing survives from all of it, let it be the order. The level is the last thing to look at, not the first. Support in a downtrend is where a falling stock pauses. It is not evidence that the fall has ended, and the premium it pays you is not evidence either.