On-Balance Volume

Jane is watching a bitcoin chart and wants to understand trading activity alongside price. Each candle summarises a time interval. Its Close is the last price and its Volume the total amount traded during that interval. Fidelity’s rule compares Close with the previous Close: a higher close adds the whole volume to the existing OBV; a lower close subtracts it. An equal closing price adds and subtracts nothing. This imaginary account is not a wallet or exchange balance. It is a way to turn trading history into another line on the chart. [1]

What the rule leaves out is revealing. TradingView’s formula uses the direction of the change in Close, not its size. If two candles have the same Volume and each closes above its previous candle, they add the same amount to OBV, even if one price rise is tiny and the other sharp. Jane cannot infer that the price moves were equally strong. Conversely, an exactly unchanged Close contributes nothing even if a lot of trading occurred. The indicator is not hiding that activity; it is applying a simple classification rule. [2]

OBV builds on its previous result, so it has to start somewhere. Fidelity recommends focusing on its direction rather than the absolute value. The running-sum calculation implies that a different starting value shifts the entire subsequent series by a constant, provided all other inputs remain the same. Jane should therefore not treat a zero crossing as a universal buy or sell boundary. Negative OBV is neither a negative cash balance nor the size of her loss. Meaningful comparison requires the same start and data, or a comparison of corresponding changes rather than raw differences in levels. [1]

Jane needs to check where the data comes from. Coinbase, for example, supplies candles for a chosen trading product and time interval, including Volume. A BTC–USD series from one exchange does not automatically combine trades across every bitcoin market. Units matter too: volume in BTC cannot simply be compared with volume in money or contract counts. OBV inherits the unit of its input volume. A consistent source, product, time grouping and unit are necessary for a meaningful comparison. Even switching the same chart to another interval can change the line’s shape, because trades and closing-price changes are grouped differently. [3]

TradingView explains OBV in terms of buying and selling pressure. That is an interpretation based on assigning volume to rising or falling candles. The formula itself contains only Close and Volume. It has no participant identities, deposits or withdrawals, nor information about whether an incoming buy or sell order initiated each trade. Every completed trade also has both a buyer and a seller. A rising line therefore cannot tell Jane how many new buyers arrived or the net capital inflow. It only tells her that positively assigned volumes dominate the running total under the chosen rule. [2]

Jane can compare price peaks with the corresponding OBV movements. Fidelity describes a price making another higher peak while OBV fails to make its own higher peak; lower price lows and the indicator’s response can be compared similarly. This disagreement is called divergence. It may be a reason to examine a trend, but it does not promise a reversal or set its date. Before testing, Jane must define which peaks and intervals she compares and when she could have recognised them. Choosing convenient points after seeing the whole chart would give the test information from the future. [1]

TradingView notes that the Close of an open candle is still changing. If the current price moves from one side of the previous Close to the other, the sign of its OBV contribution can change too, while more trades are being added. An intrabar reading is therefore different from a final value. Coinbase also warns that history may be incomplete and does not publish intervals without recorded trades. But Jane should not assume a gap means genuine zero activity without checking. She first needs to establish whether there was no trade or no record. Intrabar decisions and the treatment of gaps must be explicit in the test. [4] [3]

TradingView recommends reading OBV alongside other analysis. Even rising price and rising OBV together do not guarantee a worthwhile trade. Jane needs advance rules for entry, exit, position size, invalidation of the plan and whether she waits for the candle to close. Results must include fees and slippage, the difference between expected and actual execution prices. Repeatedly choosing rules to fit the same historical chart leads to overfitting: a rule adapts to the past but may fail elsewhere. That is why testing on data she did not use to choose the rule matters. [2] [5]