Key takeaways · 5
- New highs minus new lows nets one count against another: the number of stocks printing a fresh 52-week high minus the number printing a fresh 52-week low, on a single trading day.
- The net reading is a participation gauge. A strong positive says upside leadership is broad, a deep negative says breakdowns are spreading. Near zero, the highs and lows are just cancelling out.
- It is a different measurement from the advance/decline line. A/D counts every up day against every down day; this counts only the stocks at a 52-week extreme, the tails rather than the whole distribution.
- The warning it is built to catch is divergence: the index grinds to a new high while the net new-high count makes a lower high. Fewer and fewer names are doing the work.
- tickerstance computes it as breadth.new_highs_lows across the broad eligible universe and carries it at 10% of the Breadth subscore; the live NH − NL tile sits on the breadth page.
What new highs minus new lows counts
On any trading day, some stocks close at the highest price they have reached in a year and some close at the lowest. New highs minus new lows is the net of those two counts: the number of stocks printing a fresh 52-week high, minus the number printing a fresh 52-week low. One subtraction, computed once per session, across the whole eligible universe.
The raw number is a count, but it is usually read as a share of the universe so it travels across time. Forty net new highs means something different in a three-thousand-stock universe than in a three-hundred-stock one. Expressed as a fraction, a reading of plus two percent means fresh highs outnumber fresh lows by about sixty names in a three-thousand-name universe; minus two percent means the lows have it.
What the number measures is where the extremes are pointing. A market where hundreds of stocks are breaking out to new highs and almost none are breaking down is nothing like one where the new-low list is filling up. The net reading compresses that into a single signed number.
Net new highs minus new lows as a daily oscillator. Above zero, fresh highs dominate; below zero, fresh lows do. The middle is a washout, where the new-low list swamps the new-high list.
Why the net number matters
A cap-weighted index can be carried by a handful of giant names. New highs minus new lows cannot. Every stock at a 52-week extreme counts once, regardless of size, so the net reading is a one-vote-per-stock read on whether a move has company. When the index pushes higher and the net new-high count is strongly positive, hundreds of stocks are making new highs with it, not the three or four that dominate the index.
The reverse is the more useful read. When fresh lows start to outnumber fresh highs while the index is still near its peak, the internals are deteriorating under a calm surface. The new-low list is where damage shows up first, because a stock at a fresh 52-week low has already broken whatever support held it. A rising new-low count is the market telling you the number of broken charts is growing, whatever the headline says.
This is why the schools that watch breadth at all watch new highs and new lows. Stan Weinstein names the new-highs-minus-new-lows differential as one of his primary weight-of-evidence tools, alongside the advance/decline line and the percentage of stocks above their moving averages. It stays useful because the count is blunt and hard to game: how many stocks are actually leading or breaking.
Divergence: when the index and the list disagree
What new highs minus new lows is built to catch is divergence. An index makes a new high. You would expect the count of stocks at new highs to confirm it. Often, near the end of an advance, it does not. The index grinds a few points higher on the strength of its largest members while the net new-high count makes a visibly lower high than it did on the prior push. Fewer names are coming along.
That gap is the participation thinning out, and it does not time anything. A thinning new-high list was one of the clearer tells before both the 2000 and the 2007 tops, and it was visible in 2021 as the new-high count faded while the cap-weighted index ground on. In each case the warning was early, sometimes by many months, and acting on it cleanly was harder than spotting it.
The figure below is the schematic: price makes a marginally higher high, the net new-high count makes a lower one. The reading is not sell. It is that the advance is being carried by fewer stocks than it was, which is a condition, not an instruction. A divergence can persist for a long time, and it can resolve by breadth catching up rather than price rolling over. It is a yellow light.
The classic breadth divergence: price makes a higher high while the net new-high count makes a lower one. Fewer names confirm the move. A yellow light, not a red one.
How it differs from the advance/decline line
New highs minus new lows is easy to confuse with the advance/decline line, and they are not the same measurement. The advance/decline line counts every stock that closed up against every stock that closed down, every session. It is a read on the whole distribution: was today broad or narrow. New highs minus new lows ignores the middle of the distribution entirely and counts only the tails: the stocks at a fresh one-year extreme.
The two answer different questions. On a quiet day where most stocks drift slightly higher, the advance/decline reading is positive but the new-high count can be near zero, because drifting up is not the same as breaking to a new high. In a sharp washout both go negative, but the new-low count spikes hardest, because a fresh 52-week low is a higher bar than simply closing down. New highs minus new lows is the more selective gauge: it only fires on stocks doing something decisive.
Both belong on a breadth dashboard for that reason. The advance/decline line is the broad participation read; new highs minus new lows is the leadership-and-damage read at the extremes. Read them together. Either one alone will mislead you eventually, which is the advance/decline line essay's point too.
Where the read misleads
Universe composition. The count depends entirely on which stocks are eligible. A universe stuffed with low-priced micro-caps or rate-sensitive bond proxies will print new lows on a yield move that has nothing to do with the equity market, the same contamination the advance/decline literature warns about. A 52-week-extreme count is only as clean as the universe behind it.
Both lists long at once. Occasionally new highs and new lows are both elevated, the market splitting into a group ripping higher and a group breaking down at the same time. That churn is the basis of the Hindenburg Omen and similar signals that try to read an internally fractured market. The underlying observation is real; the record of trading those signals is poor. Note the condition, do not bet the account on the omen.
Thin history and young names. A stock needs roughly a year of trading before it can make a 52-week high or low at all, so a universe full of recent listings undercounts both. The number means less right after a wave of new issuance.
Regime. Like every breadth read, new highs minus new lows is most useful in context. New lows expanding in an already-defensive market confirms what you knew; new lows expanding while the index is calm is the more interesting message. The net count is one lens, and it reads best against the others, which is what the Stance score is for.
How tickerstance uses it
The breadth.new_highs_lows signal computes the net reading across the broad eligible US equity universe (close at or above five dollars, with a dollar-volume floor). For each stock with at least a year of history it checks whether today's high exceeds the highest high of the prior 252 sessions (a new 52-week high) or today's low undercuts the lowest low (a new 52-week low), then nets the two counts and divides by the eligible universe. The output is a signed share, normalized so that roughly plus or minus five percent net marks the strong ends of the scale.
It carries a 10% weight inside the Breadth subscore, which is itself 30% of the headline Stance score. That puts it alongside the percentage of stocks above their 50- and 200-day moving averages, distribution days, the advance/decline ratio, the McClellan Oscillator, follow-through days, and the up-down volume ratio. No single breadth signal is the read; the net new-high count is one ingredient in a deliberately diversified blend.
On the dashboard the live reading shows up as the NH − NL tile on the breadth page, banded from washout to expansion. tickerstance reports the count and where it sits; it does not tell you that an expanding new-high list means buy or that a filling new-low list means sell. The number is a condition. What you do with it is yours.
Frequently asked questions
What is the new highs minus new lows indicator?
It is a market-breadth gauge that subtracts the number of stocks making a fresh 52-week low from the number making a fresh 52-week high on a given trading day. A positive net reading means upside leadership is broad; a negative reading means more stocks are breaking down than breaking out.
How is net new highs minus new lows calculated?
Count every eligible stock that printed a new 52-week high today, count every one that printed a new 52-week low, subtract the lows from the highs, and divide by the size of the eligible universe to express it as a share. tickerstance uses the high and low of the prior 252 trading sessions as the 52-week reference.
What is a healthy new highs minus new lows reading?
A sustained positive net reading, with new highs comfortably outnumbering new lows, is the healthy backdrop for holding leaders: the advance has broad participation. A reading that turns and stays negative, with the new-low list filling up, is the warning that breakdowns are spreading beneath the index.
How is new highs minus new lows different from the advance/decline line?
The advance/decline line counts every up stock against every down stock each day, the whole distribution. New highs minus new lows counts only the stocks at a 52-week extreme, the tails. A day can be broadly positive on advance/decline while almost no stocks make new highs, so the two are related but measure different things.
What does a new-high/new-low divergence mean?
A divergence is the index making a new high while the net new-high count makes a lower high than it did on the prior advance. It means fewer stocks are participating in the move. It is a yellow light, not a sell signal: divergences can run for months and can resolve by breadth catching up rather than price rolling over.
Where does tickerstance show new highs minus new lows?
As the NH − NL tile on the breadth page, and as the breadth.new_highs_lows signal inside the Breadth subscore (10% weight). The Breadth subscore is 30% of the headline Stance score; the full math is on the methodology page.