Why Do Perpetual Contracts Experience Timed Fluctuations Every 15 Minutes?

Author: Andjela Radmilac, cryptoslate

Compiled by: Luffy, Foresight News

At 14:59:59 UTC, the price of Bitcoin perpetual contracts, like other electronic trading markets, fluctuates continuously, with orders from global traders pouring in. However, at the exact moment the clock strikes 15:00:00, market activity surges dramatically. The volume of executed orders skyrockets, and funds change hands rapidly, with significant price volatility occurring within the next 10 seconds, despite the absence of any breaking news driving trading behavior.

This trading pulse occurs predictably at 15, 30, and 45 minutes past the hour. Smaller-scale similar movements can also be observed at 5-minute intervals and at the start of each minute, but the trading explosion is strongest at the top of the hour. The crypto market is inherently a perpetual trading market, yet trading software divides it into countless micro trading periods.

Korean policy researcher Chan Kim and Peter Reinhard Hansen from the University of North Carolina documented this phenomenon in an academic paper. The research sample was drawn from the tick-by-tick transaction records of six major perpetual contracts on Binance from January 1, 2021, to October 31, 2024, covering Bitcoin, Ethereum, XRP, Solana, Dogecoin, and Cardano, spanning 1,400 complete uninterrupted trading days.

The study focuses on perpetual contracts, where traders can bet on asset price movements using leverage to amplify their positions. Traditional futures contracts have a fixed expiration date, while perpetual contracts can be held indefinitely as long as the trader's margin is sufficient; both long and short positions periodically exchange funding fees to ensure the contract price closely tracks the spot index. When the perpetual contract price is above the spot index, long traders must pay funding rates to short traders; conversely, if the contract price is below the spot index, short traders pay fees to long traders.

Perpetual contracts account for a significant share of global crypto trading, giving this short-term pulse trading a broader market impact. The prices of perpetual contracts guide arbitrage, hedging, and market-making activities, meaning that price fluctuations in the futures market will transmit to the spot market for Bitcoin and other assets.

The Crypto Market Has Its Own "Opening Bell"

If we visualize an hour as a circular chart, the 15-minute trading pulses become clearly visible. The researchers' chart shows peaks at the 0, 15, 30, and 45-minute marks, with trading volume and price fluctuations exhibiting a star-shaped distribution, where the majority of surges concentrate within the first 10 seconds after the cycle begins.

Aggregating data from the six contracts, the number of transactions in this 10-second window is 26% higher than during normal periods, with dollar-denominated trading volume increasing by 32%, and absolute price volatility expanding by 26%. The absolute return metric measures the bidirectional price fluctuation space, with both upward and downward movements becoming more pronounced at the start of the 15-minute window.

The polar chart illustrates the absolute returns and trading volume patterns of BTC, ETH, XRP, SOL, DOGE, and ADA perpetual contracts at different minute nodes within an hour.

Regardless of the market capitalization differences among the assets, this pattern holds true. During the sample period, Bitcoin averaged 1.54 million transactions per day, with a contract trading volume of $14.58 billion; meanwhile, Cardano averaged about 290,000 transactions per day, with a trading volume of only $544 million, yet both exhibited a highly consistent trading rhythm.

This cross-asset consistency is the most important conclusion of this study: this trading pattern is generated by a universally applicable trading mechanism across the market, not a unique characteristic of individual tokens.

The vast majority of trading software organizes uninterrupted price data streams into standardized candlestick charts of 1 minute, 5 minutes, 15 minutes, etc. A 15-minute candlestick summarizes the opening price, closing price, highest price, and lowest price within the period. This visualization tool not only facilitates manual observation of the market but also provides standardized data units for quantitative programs.

At the end of each candlestick, various technical indicators are recalculated, and automated trading strategies refresh trading instructions based on the latest completed candlestick. Algorithms that break down large orders will issue remaining orders at time nodes; market makers will adjust quotes based on anticipated fund flows; faster quantitative systems will position themselves in advance.

Once enough programs share the same time scale, the originally data-display-only time periods will ultimately become part of the market itself. At the end of a mundane 15-minute period, trading activity surges as if the market were opening in a traditional exchange. The real opening bell in traditional markets triggers an explosion of orders because traders have been waiting for a long time during the market closure to place concentrated orders. The crypto market, relying on unified candlestick periods and default software settings, replicates this trading frenzy, cycling every 15 minutes without stopping.

Traces Left by Algorithmic Trading

Binance's transaction records can display the trading asset, transaction quantity, and price, but they cannot distinguish whether an order comes from a human trader, a market-making institution, a liquidation engine, or other automated programs. Kim and Hansen sought indirect clues by examining the order sizes.

Human traders tend to prefer neat integer price levels and quantities, such as 0.1 Bitcoin or an order amount of about $10,000, and do not intentionally calculate a long string of fragmented trading values. In contrast, quantitative algorithms typically calculate order sizes based on volatility, available funds, current exposure, or large order splitting targets, resulting in transaction amounts that often appear random from a human perspective.

The researchers analyzed the frequency of orders ending in 0: during the few seconds when the pulse trading begins, the proportion of neatly sized orders significantly decreases. The statistical sample only included orders large enough to avoid data bias from the exchange's minimum order size, preventing small orders from being misclassified as non-human trades.

The more critical the trading node, the more pronounced the decline in the proportion of integer orders. At ordinary minute start times, the proportion of neat orders slightly decreases, while at 5-minute nodes, the decline is more pronounced, and at 15-minute nodes, it drops further, with the differential characteristics at the top of the hour reaching a peak.

For example, in the case of Bitcoin orders meeting the double-zero statistical condition: during ordinary minute opening periods, the proportion of integer orders deviates from the normal value by 0.04 standard deviations; at the top of the hour, this deviation reaches 0.20, indicating that the impact of the top-of-the-hour effect is five times that of ordinary periods.

The chart shows the trading activity of six cryptocurrencies peaking in the last 10 seconds of each minute, with higher average trading volumes at the 15-minute nodes.

Standard deviation represents the extent to which observed values deviate from the normal range; this value does not directly equate to the proportion of machine orders. It merely demonstrates that at the moment of heightened trading activity, the market no longer habitually submits integer orders, a behavioral characteristic that confirms an increase in automated trading participation.

Relying solely on order size still cannot determine the source of each order. Large institutional split orders, forced liquidation orders, and funding rate arbitrage orders can also produce irregular transaction amounts. Therefore, the paper only uses order characteristics as indirect evidence of quantitative trading activity.

The authors conducted multiple control experiments to rule out the possibility that other periodic events caused the pulse trading. During the sample period, Binance settles funding rates at 0:00, 8:00, and 16:00 UTC; even after excluding these three time windows, the 15-minute pulse effect remains significant; even when all top-of-the-hour observation data is excluded, the trading characteristics at 15, 30, and 45 minutes still exist. Independent analysis of Bybit exchange data also yielded highly similar market patterns.

The control experiments confirm that this is a widely existing electronic collaborative trading behavior. Traders could customize any trading period, but exchange data, chart parameters, and common technical indicators guide many trading programs to lock in the same time boundaries. The highest concentration of funds occurs at the most focused time nodes, such as the top of the hour and the quarter-hour.

Predictable Price Signals, Yet Insufficient to Cover Trading Fees

After confirming the periodicity of pulse trading, the research team further validated whether the market data before the opening of the 15-minute window could predict the price direction within the first 10 seconds of the opening.

The rolling prediction model reads the previous 15-minute return data, combines it with classic volume-price indicators, and uses available market information at the time to conduct out-of-sample predictions.

Backtesting results for the six contracts show that the model's accuracy in predicting price direction reaches 56.6%; the mean out-of-sample goodness of fit is 3.4%, meaning the model can only explain a small portion of the return fluctuations in those 10 seconds; the area under the curve score is 0.60 (0.5 represents random guessing, and 1.0 indicates perfect prediction).

In a high-noise market at the 10-second level, these limited values demonstrate that this trading pattern indeed possesses repeatable signal value. However, this pattern cannot be transformed into a simple and stable profit strategy, as the price fluctuations resulting from predictions are minimal. Strictly following the model signals and executing trades at each 15-minute node yields an average gross return of only 0.51 basis points before deducting fees, equivalent to 0.0051%; for a $10,000 principal, the gross profit is about $0.51.

During the sample statistical period, Binance's taker fee is 5 basis points, and the maker fee is 2 basis points. A $10,000 taker order incurs an opening cost of about $5, and closing it requires paying fees again; the average gross profit from the model does not even cover one-tenth of the single opening fee.

The profit margin is extremely thin, and the core value of this dataset reveals the vast gap between statistical predictability and the actual profits that ordinary traders can capture. Market patterns can be repeatedly validated through rigorous data testing, but short-term volatility returns are insufficient to cover basic trading costs. This is also the core reason why highly automated markets exhibit identifiable trading patterns yet struggle to easily arbitrage.

Market makers and large institutions can still leverage the conclusions of this study to optimize trading strategies. Institutions providing liquidity through bilateral orders can widen the bid-ask spread during the 10-second burst of pulse trading; when anticipating one-sided fund inflows, they can lower order amounts. Institutions executing large split orders can avoid placing orders during congested time nodes to reduce slippage losses caused by their own orders.

The market signals contained within the first 10 seconds of the 15-minute window also hold longer-term market signals. If the volume of proactive buy orders exceeds that of sell orders within the quarter-hour node, this order imbalance often continues to push prices higher in the next 4-12 hours; conversely, when sell orders dominate, the medium to long-term market faces pressure.

Order imbalance represents the difference between the volume of proactive buy and sell transactions relative to the total trading volume, used to measure which side of the market has a stronger funding offensive.

At the 4-hour dimension, the medium to long-term market largely inherits the funding flow signal from the previous 15-minute node; extending to 8-hour and 12-hour dimensions, traditional volume-price indicators provide stronger explanatory power. This market logic confirms that quantitative programs will use 15 minutes as a unified signal node to digest the long-term accumulated trading information across the entire market.

This long-term conclusion should be viewed with caution. The 4-hour, 8-hour, and 12-hour return observation windows overlap, and a single large wave of market activity may repeat across multiple sets of statistical data. Although the authors used a block bootstrap method suitable for non-independent data to handle biases, the aggregated transaction records still cannot distinguish whether the orders contain private information, whether they respond to the same public news, or whether price fluctuations merely stem from market makers absorbing large one-sided orders.

Setting aside complex statistical models, the underlying logic is straightforward. The crypto market has eliminated the closing bell, achieving continuous trading around the clock; however, API interfaces, candlestick periods, and automated trading strategies have constructed countless micro opening moments throughout the day.

Every 15 minutes, thousands of independently operating trading programs converge at the same time node. Within a matter of seconds, this market, designed to be uninterrupted, resembles a large group of traders rushing through the same door.

This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.