Where the numbers come from
An indicator is a formula applied to prices that are already on your screen. A moving average averages closes you can see. A momentum reading compares a close with an earlier close. A band is built from a mean and a spread of the same series.
That is the whole mechanism, and it has a consequence people skip past: an indicator adds no new observation. It cannot. Nothing entered the chart that was not there — the values were rearranged into a form that is easier to compare across time.
This is not an argument against indicators. Rearranging is genuinely useful; a number is easier to compare than a shape. Part 09 is devoted to what each instrument does and where it breaks. This lesson asks a narrower question: how many of them belong in your saved workspace.
Agreement between two things built from one thing
Most widely used indicators take the same input: the closing price, over a window of recent periods. Change the window and you have a different-looking line drawn from the same numbers.
So when two of them agree, they have not confirmed each other. They have re-stated one input twice, at slightly different lengths. Three agreeing is one piece of evidence counted three times, and it arrives with three times the confidence.
This is the same error 03.01 warned about with two raw rupee lines on one chart: the picture invites a comparison the underlying data does not support. Here the invitation is stronger, because agreement between independent sources genuinely is evidence — and these sources are not independent.
Add instruments, count readings
Each indicator can read up, down or neither. Add them one at a time and watch what grows and what does not.
The count is arithmetic, not a claim about any particular indicator. Some take the high and low, or volume, and are less redundant than this — the question is whether you know which yours takes.
What grows is the number of stories
With one indicator there are three readings: up, down, or neither. With five there are 243 combinations, and on an ordinary day several of them are available at the same time — some pointing one way, some the other, some flat.
At that point the chart has stopped answering you. You are choosing which subset to notice, and you will choose the subset you already believed, without ever being aware of having chosen. Nothing on the screen will contradict you, because something on the screen agrees with everything.
The cost is not that you become wrong. It is that you become unrepeatable: the same chart, on two days, with the same prices, produces two different conclusions, and you cannot tell which of the two was the analysis.
- One indicator, three states each
- 3 readings
- Three indicators
- 27 combinations
- Five indicators
- 243 combinations
- Distinct inputs, if all five use the close
- 1
- New observations added by any of them
- 0
A moving average, a momentum reading and a band all turn positive on the same chart on the same day, and all three are calculated from the closing price. What have you learned?
The test for keeping one
Fewer is not automatically better, and a bare chart is not a virtue. An indicator that answers a question your process actually asks earns its place, and taking it off would make the decision worse, not purer.
So the test is not a count. It is three answers, and an indicator that cannot supply all three comes off the saved layout:
What question does it answer — stated as a question you genuinely ask, not as a name. What input does it transform — close, or high and low, or volume, because that is how you know whether it duplicates what is already there. What would make you remove it — the condition under which you would stop using it, decided now rather than on a day it is disagreeing with you.
The third is the one people cannot answer, and it is the one that matters. An instrument you would never remove is not an instrument. It is furniture.
Which of these is a reason to keep an indicator on your saved layout?
Not accuracy. Repeatability. A chart you can read the same way twice produces conclusions you can compare, disagree with and learn from. A chart with 243 available readings produces a conclusion every time and teaches you nothing, because it never told you anything you did not bring to it. Neither chart knows what the price will do next.
One template, then, and only what survives the three questions. The last decision in this Part is how many of these workspaces you should have — and what each one is named after.