A moving average is the mean of the last n closing prices, recalculated as each new period arrives. Plotted under the price it produces a line that follows the general direction while ignoring individual candles. The window length is the only real choice: short windows track price closely and change direction often, long windows are steadier and slower.
The simple moving average weights every period in the window equally. The exponential moving average weights recent periods more heavily, so it turns sooner after a sharp move — at the cost of reacting to more moves that turn out to be noise. Neither is better in general. They answer the same question with different assumptions about how much the recent past should matter.
The lag is not a flaw to be tuned away; it is what the average is. An average of the last two hundred closes cannot signal a reversal until enough new closes have arrived to move it, which means it will always confirm a change of direction after the fact. Shortening the window reduces the lag and increases the number of false turns. This trade-off has no solution, and any system claiming to have removed the lag has moved it somewhere else.
Common uses are three. As a trend filter: price above a rising long-window average describes an uptrend, and the description is more robust than reading the candles directly. As dynamic support or resistance: widely watched averages attract reactions partly because a great many participants are watching the same line. And as crossovers, where a short average crossing a long one is treated as a signal — the best known being the fifty-day and two-hundred-day pair, whose crossings the financial press names.
Crossover systems are worth being clear-eyed about. In a sustained trend they keep a position on the right side of it, which is where their reputation comes from. In a range they whipsaw, generating a series of small losses as the two lines cross back and forth. Their published win rates are usually calculated without slippage and fees, and the difference is not trivial.
In crypto specifically, the widely quoted windows come from equities and carry no special authority. A two-hundred-day average is meaningful chiefly because so many participants watch it, which is a real effect and a different one from statistical significance. Twenty-four-hour markets also make the daily boundary arbitrary, so two providers can plot slightly different lines for the same window. Treat the average as a description of trend, not as a rule.