Price structure
Highs and lows: the two points everything else is built on
Every trend line, level, range and pattern you will ever draw is built out of swing highs and swing lows. They are simple to define and they have one property almost nobody says out loud: you cannot identify one at the moment it happens.
What a swing high actually is
A swing high is a bar whose high is higher than the highs of some fixed number of bars on either side of it. A swing low is the same definition upside down: a bar whose low is lower than the lows of the bars either side.
Notice what that definition contains. It contains a number - how many bars either side - and that number is your choice, not a property of the market. Three bars either side is common. So is five. The market does not know which you picked.
Here are seven bars, written as their highs, with a three-bar rule:
bar 1 2 3 4 5 6 7 high 101 103 105 108 106 104 103 bar 4 high = 108 the three highs before it = 101, 103, 105 all lower the three highs after it = 106, 104, 103 all lower so bar 4 is a swing high.
That is the whole definition. There is nothing predictive in it and nothing clever. It is a local maximum of a short stretch of bars, and its only job is to give you a fixed price you can refer to later.
It needs bars on BOTH sides, and that is the whole problem
Read the example again and ask the only question that matters: when did you know? Bar 4 printed its 108 and at that moment it was just a bar. You could not call it a swing high, because the rule requires three lower highs after it and those bars had not happened yet.
You learn that bar 4 was a swing high at the close of bar 7. Suppose bar 7 closes at 103.0:
the swing high = 108.00 price at the close of bar 7 = 103.00 distance = 108.00 - 103.00 = 5.00 as a percentage = 5.00 / 108.00 = 4.63% the level is confirmed 3 bars late and 4.63% away.
Nothing has gone wrong. That is arithmetic, not bad luck: a rule that asks for three bars of evidence after the event can only ever pay out three bars after the event. Every structural level on your chart was confirmed after price had already left it.
Change the rule and the pivots change
Keep the seven bars above and add one more. Bar 8 prints a high of 109. Under the three-bar rule bar 4 is still a swing high - it was confirmed at bar 7 and confirmation is not taken back. Under a four-bar rule it never was one:
bar 1 2 3 4 5 6 7 8
high 101 103 105 108 106 104 103 109
three-bar rule: bars 5,6,7 are all below 108 -> bar 4 is a pivot
four-bar rule: bar 8 is 109, which is ABOVE 108
-> bar 4 is NOT a pivot, and never becomes oneTwo analysts looking at identical data now disagree about whether there is a level at 108, and both are right, because they answered two different questions. This is the first place structure stops being objective: the data is objective, the rule applied to it is a choice.
The practical consequence is not that one rule is better. It is that you have to fix the rule before you look, exactly as the moving averages lesson says about periods. Choosing the swing rule that makes the nicest picture on the chart in front of you is curve fitting with a different name.
Wicks, closes, and which number to write down
A bar has four prices - open, high, low, close - and a swing high can be defined on the high or on the close. They are different levels and they behave differently.
- Highs and lows include wicks, so they mark the furthest price anyone actually traded at. That is the real extreme, and in thin crypto markets a wick can be a long way from everything else that happened in the bar.
- Closes ignore wicks, so a close-based pivot is more stable and sits inside the range. It also throws away the information that somebody traded out there.
- Mixing them is how a level drifts. If you mark the high and later ask whether price "broke" it using closes, you are comparing two different measurements and the answer depends on which one you happened to use that day.
Neither is correct. Write down which one you use and use only that one. The cost of being inconsistent is that you can always find a version of the chart that agrees with you, and that is never information.
What swing points are actually good for
Given all of the above, it is fair to ask what a late and rule-dependent point is worth. Three real things:
- They are fixed. Unlike a moving average, a swing high stays at 108 forever. A reference price that does not move is genuinely useful for describing what happened afterwards.
- They mark where people changed their minds. Price stopped going up at 108 because buying stopped or selling arrived. That is a fact about behaviour at a price, and it is the raw material for everything in the rest of this module.
- They are countable. Because the definition is mechanical, you can count swing points and compare them - which is what the next lesson does to define a trend without using an opinion.
What they are not is a trigger. "Price is near the last swing low" is a description of where you are, and it arrives with a built-in delay you can measure in bars. Treat it as location, never as timing.
The terms, defined
- Swing high
- A bar whose high exceeds the highs of a fixed number of bars on both sides of it. The number is a setting you choose.
- Swing low
- The mirror image: a bar whose low is below the lows of a fixed number of bars on both sides.
- Pivot
- Either of the above. A local extreme, identifiable only once price has moved away from it.
- Lookback / lookforward
- How many bars either side the rule requires. The lookforward half is what makes a pivot late, and by exactly that many bars.
- Confirmation
- The moment the bars on the right-hand side have printed and the pivot meets the rule. Always after the extreme itself.
- Wick
- The part of a bar outside its open and close. It records a price somebody traded at, which is why high-based and close-based pivots differ.
What this means in practice
- Fix your swing rule - three bars, five bars, whichever - before you open a chart, and keep it across every coin you look at.
- Decide once whether you mark highs and lows or closes, and never compare one against the other.
- Every time you mark a pivot, say out loud which bar confirmed it. That bar, not the extreme, is when the information existed.
- Multiply the lookforward bars by your timeframe to get the delay in hours. That number is how stale your structure is at best.
- Treat a pivot as a fixed reference price, not as a reason to act now.
- If a pivot only exists under one swing rule and vanishes under another, it is not a strong feature of the chart.
Nothing in this lesson is financial advice or an instruction to trade. See the risk disclaimer. We never recommend a leverage setting - that decision is yours.