The Battle of Fees and Metrics: The True Dilemma of the Squeeze Momentum Strategy and the Breakthrough of the ADX Indicator

A yearly backtest of high-frequency cryptocurrency trading strategies has shattered many traders’ dreams. Testing the popular TradingView strategy Squeeze Momentum on BTC and ETH with an initial capital of $100,000 shows that trading fees are far from just simple transaction costs—they are the primary variable determining a strategy’s survival. To break through this bottleneck, introducing trend filtering tools like the ADX indicator has become an essential solution.

How a 0.04% fee difference can flip a strategy

Data is the most convincing. Under the same strategy and market conditions, a mere 0.04% difference in fees (from 0.02% maker fee to 0.06% taker fee) can completely reverse the backtest results of ETH on the 15-minute timeframe—from a profit of 47.34% to a loss of 13.81%.

This is not marginal; it’s a fatal blow.

Specifically, the backtest set three typical fee scenarios: 0% as an ideal baseline, 0.02% representing maker order fills, and 0.06% representing taker market fills. On the 15-minute timeframe (which generates 600–800 trades annually), these different fee structures produced starkly contrasting results:

BTC 15-minute performance:

• No fees: +21.47% profit
• 0.02% maker fee: -14.45% loss
• 0.06% taker fee: -55.94% huge loss

ETH 15-minute performance:

• No fees: +68.66% explosive profit
• 0.02% maker fee: +47.34% steady profit
• 0.06% taker fee: -13.81% loss

The most ironic part of ETH’s data is that a cumulative fee of $76,536 can push a strategy that should be highly profitable into negative territory. This means that for high-frequency strategies executing over 600 trades annually, each trade implicitly carries an “entry ticket” cost of about $92.

The true cost of trading frequency at 15-minute intervals

The 15-minute cycle is where Squeeze Momentum signals are most frequent and where fee effects are magnified. Under this timeframe, BTC’s predicament is especially evident.

Even at a relatively low fee rate of 0.02%, BTC on the 15-minute chart still shows a -14.45% loss. This reflects BTC’s relatively low volatility (Beta) which makes it difficult to cover the fixed costs of frequent trading. In other words, under current market conditions, simple breakout strategies based on technical indicators struggle to generate enough alpha on short cycles.

In contrast, ETH demonstrates a stronger volatility advantage. This higher volatility allows ETH to maintain a substantial 47.34% profit at a 0.02% maker fee. But once market orders (0.06% fee) are used, even capturing major trends, the high accumulated fees can turn the account negative.

This comparison reveals a core truth: Fee structures directly determine which assets are suitable for high-frequency strategies and which should be avoided.

Why lengthening the cycle doesn’t save this strategy

The intuitive approach is that reducing trading frequency should lessen fee erosion. But when backtesting on a 1-hour timeframe, unexpected failures occur.

At the 1-hour level, BTC suffers a severe loss of -37.33%, and ETH also underperforms (-34.49%). More strangely, even with zero fees (0%), BTC and ETH still show losses of -12.29% and -11.51%, respectively.

This points to a key issue: Default parameters may completely fail across different timeframes.

The standard Squeeze Momentum parameters (BB length 20, standard deviation 2.0) work well on 15-minute charts but start to “fail” on 1-hour charts. This is likely related to signal lag—by the time the “compression release” confirmation appears on a 1-hour chart, the trend has often already started, leading the strategy to chase high points and then face deep corrections.

Introducing ADX and multi-timeframe resonance filtering for invalid signals

To fundamentally improve this situation, merely optimizing fee structures isn’t enough. The key is to reduce frequent entries during sideways or choppy markets.

The ADX indicator (Average Directional Index) plays a crucial role here. ADX measures trend strength, not direction, with values ranging from 0 to 100. When ADX > 20, it generally indicates a clear trend; when ADX < 20, the market is in a non-directional, choppy state.

Proposed improvements to the Squeeze Momentum strategy include:

Trend strength filtering: Before generating signals on the 15-minute chart, add a condition that ADX > 20. This can effectively prevent repeated entries during sideways markets, significantly reducing invalid trades and the associated fee costs.

Multi-timeframe resonance confirmation: Before opening a position on the 15-minute chart, confirm that higher timeframes like 1-hour or 4-hour also show ADX > 20 and that Bollinger Bands are in a bullish alignment. This multi-timeframe resonance greatly enhances signal validity, making each trade more profitable and increasing the value of the fees paid.

By applying these ADX-based filters, the frequent opening of positions in sideways markets can be greatly reduced, improving profitability under the same fee conditions.

Can parameter optimization break through the fee ceiling?

Besides introducing ADX, adapting parameters for different timeframes is also critical.

The original default parameters (BB length 20, std dev 2.0) are optimized for specific timeframes and may not suit others. The failure at 1-hour suggests that “one-size-fits-all” parameter settings are ineffective in high-frequency trading.

Optimization directions include:

• Fine-tuning parameters on the 15-minute timeframe to maintain trend sensitivity while reducing false signals
• Re-evaluating parameters on the 1-hour timeframe, such as increasing BB length to 30–40 to reduce lag, combined with ADX confirmation
• Dynamically adjusting parameters to suit different market cycles, avoiding being caught in low liquidity or high volatility periods

However, parameter optimization can only marginally improve returns; it cannot break through the hard ceiling imposed by trading fees. The real breakthrough depends on fee structure itself.

From backtest to live trading: survival rules for high-frequency strategies

This backtest ultimately points to an unavoidable reality: In live trading, fees are more decisive than indicator selection.

For traders aiming to deploy similar high-frequency strategies, the key recommendations are:

First, prioritize Maker (limit) orders. Implement passive order placement logic at the algorithmic level, placing orders at the best bid or within the order book to wait for fills, rather than market taker orders. If the exchange’s taker fee exceeds 0.05%, such high-frequency strategies become practically infeasible.

Second, favor assets with higher volatility (higher Beta). Compared to BTC’s low-volatility “asset-like” profile, ETH and other major altcoins offer better cost-benefit ratios. Under fixed fees, higher volatility makes it more likely that strategies generate excess returns sufficient to cover costs.

Third, use trend strength tools like ADX for filtering. Avoid repeatedly trading in sideways markets; only open positions when a trend is confirmed (ADX > 20) and multi-timeframe resonance exists. This significantly improves trade quality.

Fourth, adapt parameters according to market conditions. Different environments and timeframes require different settings. Regular backtesting and parameter tuning are essential.

The biggest lesson from this backtest is that success in quantitative trading never relies on discovering “magical indicators,” but on meticulous control of execution costs, deep understanding of market structure, and rational filtering of signals with tools like ADX. Code optimization is just the first step; optimizing fee tiers, choosing appropriate execution methods, and ensuring liquidity often have a greater impact on final returns than parameter tuning alone.