A trendline is a drawing made from past prices. That makes it useful as a compact description of direction and potential support or resistance. It does not make it a forecast.
The useful question is not whether a chart can be made to look orderly. It is whether the exact way you draw, enter, exit and pay to trade a trendline still holds up when the next sample is unseen.
What a trendline actually is
A trendline joins selected pivot highs or pivot lows and extends that relationship across the chart. Educational references describe it as a visual tool for the prevailing direction of price, often used to frame support and resistance (Investopedia: trendline definition, IG: what a trendline is).
An ascending line is normally drawn beneath a sequence of rising lows. A descending line is normally drawn above falling highs. Those are descriptions of the points chosen; the line does not contain an independent reason for price to react.
Rising lows Falling highs
/ \
o / o \
o/ o \
---/---- projected line -------\--- projected lineWhere the rule becomes subjective
The line is not uniquely determined by a price series. A chartist can choose highs, lows or closes; a different timeframe changes the pivots; new prices can force a redraw. Both practical guides flag these choices and the need to adjust the line as data arrives (Investopedia: limitations of trendlines, IG: drawing and validating trendlines).
That is the central weakness of a discretionary trendline. The National Bureau of Economic Research notes that geometric shapes in historical charts can be in the eye of the beholder; its response was to make pattern recognition systematic rather than rely on a hand-drawn judgement (Lo, Mamaysky and Wang at NBER, the paper PDF).
There is nothing wrong with using a line as a chart-reading aid. But “the line looked clean” is not a reproducible entry condition. It cannot be meaningfully compared with another rule until two people, or one program twice, would produce the same signal from the same bars.
A break and a bounce are different hypotheses
Trendline methods usually turn into one of two ideas: price may continue in the direction of the line after touching it, or a crossing of the line may mark a change in behaviour. The same guides present lines as support/resistance references and discuss a breach as something traders watch; neither convention supplies a universal entry or exit rule (Investopedia: trendlines and breaches, IG: support, resistance and trendlines).
Treat them as separate hypotheses. A touch rule needs a definition of “touch”, a decision time and a protective exit. A break rule needs a definition of “cross”: intrabar, close beyond the line, or a distance adjusted for volatility. A line projected from candle lows and a break judged from a closing price may be sensible, but it must be written down before the result is known.
| If the claim is | The rule must say |
|---|---|
| Trendline continuation | Which pivots qualify, how the line is fitted, what counts as a touch, entry timing and exit |
| Trendline break | Which price crosses, whether a close is required, gap treatment, entry timing and exit |
| Trend channel | How the second boundary is constructed and how simultaneous signals are handled |
The vocabulary is familiar. The implementation is where the strategy lives.
What the evidence can and cannot say
Evidence on technical analysis is not a licence to trust every line on every chart. The NBER study found that some formally identified technical patterns provided incremental information in its historical US-stock sample, while also stressing the subjectivity that motivated its algorithmic method (NBER study and abstract, MIT-hosted paper). That is a much narrower claim than “trendlines work.”
A later study tests a large set of technical rules across developed and emerging equity markets and explicitly controls for data-snooping. Its design is a useful warning: an apparent rule must survive the fact that many alternatives could have been tried (Kellard, Review of Behavioral Finance, NBER on why systematic pattern definitions matter).
Neither paper establishes an edge for the line you have drawn today. Markets, costs, instruments, holding periods and the exact rule all differ. The honest conclusion is smaller: a trendline can generate a hypothesis, but it cannot validate that hypothesis by looking persuasive after the move.
How you'd actually test it
First remove the discretion. For example, define pivots with a fixed bar rule, fit the line from named pivots only, and permit a signal only after the line exists. Do not move a pivot because the resulting slope looks more convincing.
| Test component | Freeze it before the backtest |
|---|---|
| Market and bars | Instrument, feed, timeframe, timezone and sample dates |
| Line construction | Pivot rule, price field, minimum separation and refit rule |
| Signal | Touch or break definition, bar-close or intrabar timing, and re-entry policy |
| Risk and execution | Stop, exit, sizing, spread, commission, slippage, gaps and overlapping positions |
| Evaluation | A base rule for comparison, untouched out-of-sample data and the results to report |
Compare the trendline rule with a plain alternative that takes the same kind of exposure. Keep losing signals, skipped signals and costs in the record. Then carry the frozen design into data that did not help choose the pivots or the parameters. Out-of-sample testing, parameter sensitivity and realistic backtest costs explain the three checks that most chart rules evade.
realbacktesting is a trading-software studio for cTrader built around verifiable work: its published cBot figures are cTrader-native backtests that readers can reproduce in their own cTrader. A trendline deserves the same standard. Make the drawing rule executable, include friction, and let unseen data have a vote.
Frequently asked
Are trendlines support and resistance?
Trendlines are commonly used to visualise potential support and resistance from selected pivots. They do not prove that future price must react at the projected line, so the reaction rule remains a testable hypothesis.
How many touches make a trendline valid?
There is no universal answer. Guides use different conventions, and the selected price points are one reason trendlines are subjective. A test must declare its own pivot and validation rule before examining outcomes.
Does a trendline break predict a reversal?
No. A crossing is a defined chart event, not proof of a reversal. It needs a pre-written entry, exit and cost model, then an out-of-sample comparison with an alternative rule.
The stubborn takeaway
A trendline becomes useful when its drawing can be repeated. Until then, it is a neat explanation of the past, not evidence about the next trade.