Imagine two lines smoothing out the bitcoin price: a shorter average usually responds faster, and a longer one more slowly. When the shorter moves from below the longer to above it, an upward crossover has occurred. Recognising it requires comparing the previous and current values, rather than simply checking which line is higher now.
Jana adds a shorter and a longer average to her chart to make price fluctuations easier to follow. The shorter line was below the longer one, but after the next completed period it sits above it. That change is a Moving average crossover. If it remains above afterwards, that is not another crossover. The distinction matters: a rule to buy on a crossover should not repeat the purchase simply because the lines keep the same order. [1]
The shorter average usually adjusts to new prices more quickly, while the longer one retains more of the earlier movement. Their meeting therefore shows a change in the relationship between shorter-term and longer-term averages. Yet both lines may use the same price history. They are not two independent pieces of evidence that prices will rise: they process information already available and lag behind price movements. [2]
SMA is the ordinary average of a chosen number of equally weighted values. EMA gives newer values more weight. Both depend on the length, the price input called Source and the Timeframe: the period represented by each candle. Close means the closing price, or the latest price while the candle is still open. Equality matters too. In the example below, an upward crossover requires the fast average previously to have been no higher than the slow one and now to be strictly higher. Merely touching does not meet that rule. [2]
Jana spots a crossover during the current period, but the price then falls and the lines move back. This is not necessarily a chart error. The Close of an open candle is still changing, as are averages calculated from it. A rule that waits for the period to close can therefore produce a different result from one that reacts immediately. Comparisons with historical data must use the same timing: the finished chart does not show every fluctuation Jana saw live. [3]
Even a confirmed signal is only information. An order can then be created and a trade executed at an available price. A signal identified after a candle closes cannot be assumed to have been tradable at that candle’s Close. TradingView’s trading simulator, called the broker emulator, also distinguishes these steps. Results depend on execution settings, not just where a marker appears on the chart. [4]
A long position aims to benefit from a price rise; a short aims to benefit from a fall. In Jana’s rule, a downward crossover might mean selling bitcoin she already holds and closing her long. It need not open a short. Reversing a position is a different action again. A strategy must therefore specify existing holdings, order size and what the change in line order should actually trigger. [4]
Without a sustained direction, the averages can cross repeatedly. Frequent purchases and sales incur fees. Slippage, the difference between the expected and actual execution price, can add to the cost. Attractive chart markers are therefore not enough: what matters is the result after costs and how much trading it took to achieve it. [2] [4]
Jana can keep changing average lengths until she finds an impressive past result. But she may simply be fitting the rule to chance features of that period: this is called overfitting. A more useful check keeps the rule fixed and tests it, including costs, on a period not used for tuning. If she adjusts the settings again in response to that result, she needs another independent check. Even a successful test cannot guarantee future profit. [4]
One rule, two consecutive observations
In this fictional example, Jana uses completed values only. She counts an upward crossover when the fast average was previously no higher than the slow average and is now strictly higher. Initially the fast value is 100000 and the slow value 100010 USD/BTC: 100000 − 100010 = −10 USD/BTC, so the fast one is lower. At the next observation they are 100030 and 100020 USD/BTC: 100030 − 100020 = 10 USD/BTC, so the fast one is now higher. Their order has changed and the condition is met. If the fast line then remains above the slow line, that is not another crossover. The calculated difference describes the two lines, not Jana’s execution price or profit.
For the clearest picture, read this entry together with Moving average, Simple moving average, Exponential moving average, Backtesting, Overfitting, Slippage. The reverse links also lead from Moving average, Simple moving average, Exponential moving average, Trend Following.
01Is every point above the slow average a new signal?+
Not under a crossover rule. There must be a change between the previous and current order of the lines. Remaining above the slow average can be a separate condition in another strategy, but it is not a series of new crossovers.
02Does a downward crossover require opening a short?+
No. A rule might specify only closing a long position, which seeks to benefit from rising prices, or opening a short, which seeks to benefit from falling prices. It might also require no action. The crossover alone determines neither order size nor order type; those depend on the strategy and the existing position.