Gate Research’s latest market review describes May as a month in which the crypto market shifted from an early rebound into failed breakouts, pullbacks and low-volatility compression. The key conflict was not simply price weakness, but the divergence between spot absorption and a more leveraged trading structure. BTC, ETH and SOL all formed interim highs in the first part of the month before entering a corrective phase. BTC moved from a 4H closing price of $77,117.4 at the start of the month to $73,684.0 by month-end, a monthly return of -4.45%. ETH fell from $2,283.02 to $2,007.0, for a monthly return of -12.09%. SOL declined from $83.90 to $82.44, or -1.74% for the month. The relatively small closing decline in SOL masked a much wider intramonth path: SOL reached $98.40 at its monthly high and later pulled back toward the $80.00 area.
The report divides the month into three trading phases. From May 1 to May 6, BTC rose from $77,117.4 to $82,828.2, while ETH moved from $2,283.02 to $2,423.99. SOL’s advance continued until May 11, when it touched $98.40. During this first phase, short-cycle moving-average clusters moved upward, volatility stayed within a manageable range, and the market showed the characteristics of a repair rally. SOL led in terms of elasticity, indicating that traders were willing to take higher risk exposure at the start of the month.
The second phase began on May 7. BTC failed to hold steadily above $82,000, ETH could not maintain levels above $2,400, and SOL formed its monthly high around $98. Breakout signals then began to fail more frequently. Prices fell back toward EMA12 and repeatedly triggered exit signals. BTC long trades after May 14 hit stop-losses, ETH long trades after May 6 continued to fail, and SOL entered a clear downside leg after May 15. The third phase ran from May 22 into month-end, with BTC shifting down toward the $73,000 area, ETH approaching $2,000, and SOL returning toward $82. Volatility narrowed, moving-average-cluster width fell, and the market entered another compression state.
Asset roles were confirmed by the scale of intramonth drawdowns. BTC’s maximum drawdown from high to low was about 12.5%, while ETH and SOL recorded drawdowns of about 18.8% and 18.7%, respectively. Gate Research characterizes BTC as the market’s risk anchor, while ETH and SOL acted as amplifiers of risk appetite. Once BTC weakened, ETH and SOL declined faster, which meant that strategy-level long exposure to higher-beta assets needed to be reduced. As of May 31, total stablecoin market capitalization was about $320 billion and DeFi TVL was about $251 billion, showing no systemic retreat in underlying dollar liquidity. However, CEX 24-hour spot turnover was about $124.2 billion, while CEX perpetual turnover was about $894.4 billion, making perpetual volume roughly 7.2 times spot volume. Price discovery was therefore borne more heavily by derivatives markets.
ETF flows became an important source of pressure in the second half of May. Public news cited in the report showed that spot BTC ETFs recorded nine consecutive trading days of net outflows totaling about $2.8 billion. At one point, single-day net outflows reached about $649 million, including about $448 million from BlackRock IBIT. ETH ETFs were also under pressure, with net outflows of about $241 million in the final week of May. Derivatives data aligned with this picture: the taker buy/sell ratios for BTC, ETH and SOL were all below 1, showing slightly stronger active selling, while funding rates were around 0.01% and had not reached an extremely crowded level. In Gate Research’s description, May was a market with active leveraged trading, insufficient spot follow-through and weak active buying, so price breakouts were prone to becoming false moves.
The crypto market’s performance also needed to be viewed within the broader U.S. equity risk-budget framework. BTC’s correlation with the S&P 500 remained high. Public samples showed that the 30-day correlation in 2026 had reached about 0.74 and was still around 0.6 near the end of May. On the U.S. equity side, core support came from AI and large-cap technology stocks. Nvidia announced strong quarterly results in May, with Q1 FY2027 revenue of about $81.6 billion. Its share price at one point reached a record high and again touched an approximately $5 trillion market-cap milestone. AI heavyweight stocks supported Nasdaq risk appetite through earnings confirmation, while crypto assets lacked an earnings anchor of similar strength and were more affected by ETF flows, derivatives leverage and liquidity expectations.
This cross-asset divergence helped explain the strategy results. U.S. technology leaders had earnings-driven support, while mainstream BTC and ETH ETFs experienced outflows and weaker absorption. Capital within large risk assets was reallocated toward AI leaders with higher perceived earnings certainty, while allocation willingness toward BTC ETFs declined. As a result, crypto long breakouts lacked spot confirmation, and volume-driven upside moves often turned into rallies that faded.
Gate Research’s strategy review delivered a clear ranking. Equal-weight buy-and-hold across BTC, ETH and SOL returned about -6.09%. The long-only moving-average-cluster breakout strategy returned about -3.65%. The two-way moving-average-cluster breakout strategy returned +2.11%, creating roughly +8.2% of excess return versus buy-and-hold. The gains mainly came from short trend legs after mid-May, with ETH and SOL contributing most clearly. The effective framework for May was to first identify compression in the moving-average cluster, then allow price to choose either an upside or downside direction; EMA12 managed failed signals, a fixed 2.5% stop-loss limited single-trade losses, and a 3R take-profit, equal to 3 times 2.5% or 7.5%, preserved trend gains.
The system used six moving averages to construct the cluster: EMA6, EMA12, EMA24, SMA6, SMA12 and SMA24. Cluster width was calculated as the maximum value of the six averages minus the minimum value, divided by the current closing price. If the previous candle’s cluster width was below the 2.2% threshold and the current candle closed above the cluster’s upper boundary, the system entered long at the next 4H open. If the previous candle’s cluster width was below 2.2% and the current candle closed below the lower boundary, the system entered short at the next 4H open. Exit rules were fixed: longs exited when price broke below EMA12, shorts exited when price broke above EMA12. A single-trade stop-loss was set at 2.5%, take-profit at 3R, and if take-profit and stop-loss were triggered in the same candle, stop-loss took priority. Round-trip trading cost was deducted at 8bp, and any month-end position was closed at the final 4H closing price.
The long-only version failed overall. BTC made 11 trades, returned -5.36%, had an 18.2% win rate and a maximum drawdown of -10.08%. ETH made 10 trades, returned -6.49%, had a 10.0% win rate and a maximum drawdown of -10.64%. SOL made 11 trades, returned +0.91%, had an 18.2% win rate and a maximum drawdown of -7.11%. BTC’s long-only gains were concentrated in two early-month trades: the May 1 entry exited on May 4 with a net return of +2.09%, and the May 4 entry exited on May 7 with +0.92%. Signal quality then deteriorated, and the long entered on May 14 hit a stop-loss for a net loss of -2.58%.
ETH was the weakest long-only asset. The May 1 entry exited on May 5 with a net return of +3.17%, but the following nine long trades all lost money. Gate Research’s conclusion was that most ETH upside breakouts were weak rebounds rather than trend expansion. SOL remained slightly positive in the long-only test, but its return was highly concentrated in two trades: the May 5 entry exited on May 8 with +3.95%, and the May 8 entry hit 3R take-profit on May 10 with +7.42%. Most other SOL signals lost money, making SOL the only asset with positive long-only returns, but with gains concentrated in a small number of trades.
The two-way version improved the results materially. BTC’s two-way strategy still lost money, returning -2.83%, but the loss was smaller than in the long-only version. BTC made 18 trades, with a 22.2% win rate and a maximum drawdown of -10.74%. Two short trades contributed the most: the May 15 short exited on May 20 with a net return of +2.35%, while the May 26 short exited on May 30 with +3.42%. BTC had many false signals in mid-May, and repeated switching between long and short positions created friction.
ETH’s two-way strategy returned +3.14%, with 18 trades, a 38.9% win rate and a maximum drawdown of -8.26%. Its key trade was a May 15 short that hit 3R take-profit on May 17, producing a net return of +8.03%. A May 26 short exited on May 29 with a net return of +2.68%. ETH’s long signals failed, while short legs formed the main source of return. SOL’s two-way strategy returned +6.05%, with 22 trades, a 22.7% win rate and a maximum drawdown of -8.17%. SOL delivered both long and short trend trades: the May 8 long hit 3R take-profit at 16:00 on May 10 for +7.42%, and the May 15 short hit 3R take-profit on May 17 for +8.03%. SOL showed the strongest trend elasticity, while also carrying the highest trading noise.
Across all 58 two-way trades, the number of profitable trades was not high. BTC’s win rate was 22.2%, ETH’s was 38.9%, and SOL’s was 22.7%. Strategy returns came from a small number of large trend trades, while losses were controlled by EMA12 exits and fixed stop-losses. The cumulative trade path showed strategy equity moving up and down in early May, rising in mid-May on ETH and SOL short trades, and receiving further support in late May from BTC and SOL shorts. Losses were concentrated in phases with repeated long-short switching. The system’s profile was low win rate and high payoff ratio, suitable for markets with clear trend legs and less suitable for dense consolidation.
The long-short contribution breakdown gave the same message. BTC’s long contribution was negative and its short contribution was positive. ETH’s long contribution was negative and its short contribution was significantly positive. SOL had positive contributions from both longs and shorts, with shorts more stable. May’s core theme was the downside trend that followed failed upside attempts. In terms of win rate, expected return and maximum drawdown, SOL had the highest single-trade expectation, ETH ranked next, and BTC was the weakest. BTC had the densest false-breakout environment, ETH had a cleaner direction, and SOL had stronger elasticity.
Exit analysis showed that EMA12 exits accounted for the largest share. Many trades did not end through stop-loss; instead, they exited after failed breakouts returned to the EMA12 area. This rule reduced both holding time and loss expansion for incorrect signals. Stop-loss trades were limited in number but concentrated in loss contribution. 3R take-profit trades were few, yet contributed meaningfully to overall gains. The structure matched a trend-following strategy: most trades produced small losses or small gains, while a few trend trades generated most of the return. The report stated that removing the 3R take-profit would have cut off major SOL and ETH gains too early, while removing the EMA12 exit would have enlarged losses during choppy phases.
The report also tested an enhanced-filter version. Its conditions included: 7-day volatility not exceeding 1.15 times 30-day volatility; volume not below 0.9 times the 20-bar average, with one bar equal to four hours; longs near the 20-bar high; and shorts near the 20-bar low. This version performed worse. BTC’s enhanced two-way strategy returned -3.40%, ETH returned -5.03%, SOL returned -2.58%, and the equal-weight three-asset result was -3.63%. The reason for failure was that volume-driven breakouts in May often appeared near interim tops. BTC’s May 4 enhanced long signal entered at $80,322.9 and hit stop-loss within four hours, with a net loss of -2.58%. ETH’s May 6 enhanced long signal entered at $2,410.39 and hit stop-loss in the same candle, also losing -2.58%. SOL’s May 4 enhanced long signal also hit stop-loss.
Gate Research’s asset-level conclusions were distinct. BTC was the state anchor: its monthly decline was smaller than ETH’s and its drawdown was more controlled, but the BTC two-way strategy still returned -2.83%, showing that BTC was not the best return asset in May. ETH was the weak main line: it fell -12.09% for the month, long breakout win rate was extremely low, and two-way returns depended on short profitability. After failing around $2,400, ETH successively broke below $2,300, $2,200 and $2,100. SOL was a trading asset: its monthly closing loss was only -1.74%, but the path was very volatile. SOL’s two-way strategy returned +6.05%, clearly above BTC and ETH, making it better suited to trend following than passive holding within this review.
For June, the report keeps the two-way 4H moving-average-cluster breakout system and reduces the weight of unilateral long chasing. BTC remains the state filter, while ETH and SOL act as return assets after relative-strength confirmation. If BTC returns above EMA12 and the 30-bar moving-average area, ETF outflows slow, and the taker buy/sell ratio moves back above 1, long-signal weight can be raised. If BTC remains below the $74,000 to $76,000 area, the market stays in a weak repair state. The U.S. equity filter remains part of the framework: strong Nasdaq and AI leaders, together with slower BTC ETF outflows, represent renewed improvement in cross-asset risk budgets; strong Nasdaq but continued BTC ETF outflows shows capital preference for U.S. technology leaders with higher earnings certainty; synchronized weakness in U.S. equities and crypto raises the priority of ETH and SOL short signals.
The position rules stay mechanical: single-trade risk at 2.5%, take-profit at 3R, and EMA12 exits unchanged. Breakout signals alone should not trigger heavy positioning. When spot absorption is weak, ETF outflows persist, perpetual turnover remains high, and active buying is insufficient, upside breakout signals are downgraded and downside breakout signals are upgraded. Gate Research’s conclusion for May is that the crypto market completed a state switch from repair to failure. Buy-and-hold performed the worst, long-only breakouts could not adapt to trend failure after mid-month, and the two-way moving-average-cluster breakout system delivered the best result. In a low-volatility compression environment, disciplined two-way execution, controlled single-trade risk and preservation of trend gains proved more effective than subjective chase-buying.