How this instrument works
A simple moving average takes a fixed number of the most recent observations in a series and averages them, then slides that window forward by one as each new observation arrives, dropping the oldest reading to make room. That sliding is the entire point: a plain running average keeps every past value forever, so after a long enough series one more data point barely nudges it, while a moving average stays equally responsive to fresh data no matter how much history has piled up behind it.
Window length is a trade-off between responsiveness and smoothness. A short window, like the three values this instrument averages, tracks a shift almost as soon as it happens — useful to a retail buyer eyeing last week's sell-through or a landlord watching trailing maintenance costs. A longer window, such as the 50-day or 200-day figures technical traders chart against each other, barely moves on any single session's close, which is exactly why a shorter line crossing a longer one — a golden cross — reads as a slower, more confirmed signal than one day's price ever could.
The window's equal weighting is also its limit. Every reading inside it counts the same until the moment it drops out, so a rising average can coincide with a falling latest reading if the value being discarded was even lower than the one replacing it — a common misread. An exponential average avoids that by weighting recent points more heavily and never fully discarding old ones; this instrument does neither, and its three-slot window is fixed so the arithmetic stays checkable by hand.
- Enter your oldest reading into Value 1 (oldest) — the first observation in the three-period window.
- Enter the middle reading into Value 2, in the order it actually occurred.
- Enter the newest reading into Value 3 (most recent).
- Read the 3-period moving average — it recalculates the instant any of the three entries change.
Worked example — three trailing closing prices
A stock closes at $100, then $110, then $105 across three sessions — enter those, oldest first, as Value 1, Value 2, and Value 3. Summing gives 100 + 110 + 105 = 315, and dividing by 3 gives a three-period moving average of exactly 105.00, sitting between the low and the high and pulled slightly toward the $110 middle session.
That figure is small enough to check by eye, which is the reason to try a short window first: the same summing-and-dividing arithmetic scales unchanged to a 50-day or 200-day span a trader watches for a golden cross, or to the trailing three months of revenue a finance team averages before showing a board a trend line that one unusual month should not be allowed to distort.
Questions
How is a moving average different from an all-time running average?
A moving average keeps only a fixed number of the most recent periods and drops the oldest one as each new value arrives, so it stays equally sensitive to fresh data no matter how long the series runs. An all-time running average keeps every past value forever, so once a series is long enough, one more data point barely shifts it — exactly the sluggishness anyone watching this month's shift wants to avoid.
Why do traders quote 50-day and 200-day windows instead of a three-value one like this?
Window length trades responsiveness for smoothness. A short window such as this three-reading example reacts to new data almost immediately. A 50-day or 200-day window barely moves on any single day's close, which is exactly why traders watch for the shorter line crossing the longer one — a golden cross — as a slower, more confirmed signal than any one day's price could give alone.
Can the average rise even when the newest reading is lower than the last one?
Yes. The result reflects the three values currently inside the window, not the direction of the latest change alone. If the oldest reading being dropped was lower than the new one replacing it, the average can rise even though the newest entry fell from the prior reading — a common misread that confuses a rising average with a rising latest figure.
Is this the same calculation as an exponential moving average?
No. Here all three readings are weighted identically — the oldest counts exactly as much as the newest until it exits the window. An exponential moving average instead weights recent readings more heavily and never fully drops older ones, which reacts faster but is a genuinely different formula; check which kind a chart or report is using before comparing two figures.
Where besides stock prices does a three-period average like this get used?
Anywhere a single reading is noisy but the underlying direction matters: a retailer smoothing weekly unit sales before reordering stock, a quality-control chart averaging the last few measurements to catch drift before it breaches a tolerance limit, or a finance team reporting trailing three-month revenue so one large renewal does not read as a permanent jump. The arithmetic is identical in each case — only what 'value' represents changes.
Why does the instrument fix the window at three entries instead of letting me add more?
Three values keep every step of the sum-then-divide arithmetic visible and checkable by hand, which is the clearest way to see what any longer window — 10 periods, 50 days, 200 days — is doing under the hood: the same addition and division, just repeated over more entries and recalculated one slot at a time as the newest reading replaces the oldest.
References
Read this first: This instrument shows arithmetic, not advice. Real offers add fees, taxes and terms that vary by lender and place — verify the figures against your actual paperwork before deciding anything.