Direct answer
Swing Highs Lows uses chart inputs that come from price itself: the sequence of swing high points and swing low points that you identify on a price chart. In other words, the “inputs” are not external indicators by default; they are the selected high/low coordinates produced by your swing-point definition.
Because the term “Swing Highs Lows” can be implemented with different swing-identification rules, the only reliably correct way to answer the question is to list the inputs you must define for your version: (1) the price series you mark on (e.g., high/low values), (2) the rule for detecting swings (the material definition), and (3) the chart scope (time window and bar/tick granularity).
Mechanism and definition
A swing high is a local peak in price structure according to a chosen rule; a swing low is a local trough according to the same kind of rule. Once you have those points, “Swing Highs Lows” typically refers to reasoning about the ordering and relative positions of those points—such as whether new highs are higher/lower than prior highs, and whether new lows are higher/lower than prior lows.
To make this precise, your inputs usually fall into these categories:
- Price data: the values needed to mark highs and lows (commonly the bar’s high and bar’s low, or the relevant high/low from whatever chart data feed you use).
- Swing detection rule: a deterministic condition that decides whether a bar (or timestamp) qualifies as a swing high/low. For example, many implementations require the candidate point to be higher than surrounding points for swing highs, and lower than surrounding points for swing lows.
- Lookback/lookaround window (dependency): how many bars around a candidate are considered when judging whether it is “higher/lower than the neighbors.”
- Chart granularity and time window: the timeframe (e.g., minute vs. daily) and the span of history used to mark swings.
- Data consistency assumptions: whether you use the same session data treatment across time (e.g., continuous vs. session breaks), because it can change where swings appear.
With those inputs defined, the mechanics are simple: you produce an ordered list of swing highs and swing lows from the chart, and then any further analysis is derived from that list (for instance, comparing relative positions of successive swings).
Evidence or example (with explicit assumptions)
Assume you analyze a chart built from bars, where each bar has an H (high) and L (low). Suppose your swing rule is: a bar qualifies as a swing high if its high H is greater than the highs of a fixed number of neighboring bars on both sides; a bar qualifies as a swing low if its low L is less than the lows of neighboring bars on both sides.
Under this assumption, the “inputs” to the swing-high detection step are:
- the sequence of H values for each bar within your chosen timeframe,
- the neighbor count (your lookaround window),
- the candidate bar index.
Similarly, the “inputs” to the swing-low step are:
- the sequence of L values within the same timeframe,
- the same neighbor count logic.
Once you have the swing points, any structural comparisons you do (like comparing the relative height of the next swing high to the previous swing high) depend on those exact marked points. If you change the neighbor count or timeframe, the marked swing list can change, and the downstream conclusions change too.
Limitations and risks
A key limitation is that Swing Highs Lows is only as consistent as the swing-identification rule and chart scope you choose. Common failure modes include:
- Ambiguity near the threshold: if two nearby bars have similar highs/lows, small changes in data resolution can flip whether a point qualifies. - Inconsistent parameters: using one lookaround window for some observations and another for others can create a misleading sense of “structure. ”
- Overfitting to the past: tuning the swing rule to make historical examples look clean can reduce reliability on new charts. - Non-stationary conditions: historical relationships between swing points and subsequent price behavior do not guarantee similar behavior in the future; markets can change regime, volatility, and microstructure.