The report provides a systematic analysis of the OKX Boost mechanism based on data, aiming to connect and guide various stakeholders in the ecosystem.

For project teams, correlation analysis helps them understand the ranking logic and guides them to improve their ranking position in an organic and sustainable way, rather than relying on short-term data manipulation. For ordinary users, value flow and cost breakdown help them clarify their position in the entire value chain, helping them find their true break-even point and avoid becoming unknowingly burdened with hidden costs. For the OKX Boost product itself, trend analysis of macro data objectively presents its current operational status and potential structural risks, aiming to provide reference and suggestions for the healthy and long-term development of the product.

Key Takeaways

• The median reward per user across 68 projects is approximately $18, but the optimal path cost is already $10–13, and the medium path cost $32, meaning losses can easily occur with the slightest misstep. This implies that in the vast majority of projects, ordinary users cannot achieve stable positive returns through Boost, and are instead using slippage losses to subsidize external limited partners (LPs). In the long run, this will lead to a decline in the retention of genuine OKX Boost users, with the remaining participants shifting towards professional "wool-gathering" (exploiting) tactics and bots, accelerating the degradation of data quality.

• The median of $499K versus the mean of $901K indicates that a very small number of large projects (Plasma $10.7M, PerpStock $8M) have significantly inflated user expectations, while most projects can only offer reward pools of $200–700K, and the average real return per person for the vast majority of projects is only between $10 and $30.

• Data shows that OKX Boost had an average of about 45,000 participants in its early stages (September 2025), but experienced a sharp drop between November and December 2025, before stabilizing at around 20,000 in the following months (early 2026). The latest average participation data for OKX Boost is surprisingly close to $800, indicating that the majority of OKX Boost users are now semi-professional users who frequently use the platform for promotional activities.

• OKX Boost's ranking mechanism filters token trading activity with genuine market depth, primarily considering trading volume, liquidity, and the number of unique addresses. Liquidity is the threshold ($500K), while market capitalization, number of tokens held, and price fluctuations are all pseudo-correlated—this indicates that OKX is intentionally excluding long-tail assets with many holders but no trading, focusing on projects that can generate genuine price discovery on DEXs.

1. What is OKX Boost?

Essentially, it is a points/rewards ecosystem jointly launched by OKX Wallet and OKX DEX aggregator. Its core goal is to directly connect real active users with high-quality on-chain projects, while providing projects with a clear potential path from "organic on-chain growth → Boost leaderboard → OKX CEX spot listing".

Two core models :

• X Launch : Users share the project's token prize pool based on Boost data (currently the main gameplay).

• X Campaign : The project holds trading competitions from time to time, with top-ranked participants sharing the prize pool.

1.1 OKX Boost Ranking Rules

All event eligibility and reward allocation are based on Boost data from the past 10 days , updated daily at 00:00 UTC with the previous day's data.

Boost trading volume (average over the past 10 days, generated through the OKX DEX aggregator)

• Only transactions on the Solana + EVM compatible chain on the OKX DEX aggregator are recognized (cross-chain, CEX, third-party DEX such as Uniswap paths, and API transactions are not counted).

• Calculation formula : Actual transaction amount × multiplier (× Boost leaderboard bonus)

• Trading Pair Multiples Table (Official Core Rules):

1.2 Analysis of Project Ranking Logic

Form + BD → X Launch (usually lasts 24 hours) → Tradeable on OKX DEX (enters Boost Ranking) → Consistently ranked in the top 10% → Listed as a Pre-market perpetual contract → Converted to a standard perpetual contract upon Spot listing. Simply listing on Boost is meaningless for a project; to gain perpetual profits, it's necessary to consistently rank in the top 10%.

Boost Ranking's token score combines four dimensions: DEX popularity rankings, market capitalization, liquidity, and community sentiment. We analyzed 24-hour data from 46 OKX Boost projects on the list and then inferred the possible factors considered in the ranking.

As shown in Figure 1, the horizontal axis represents the correlation coefficient; the further to the left (the more negative), the larger the indicator and the higher the ranking. The conclusion is that the correlation coefficients for transaction volume, liquidity, and the number of unique addresses are all between -0.75 and -0.79, making them the main predictive signals for ranking.

• Trading volume (r = −0.793, true signal). The most correlated metric, remaining significant even after removing the influence of liquidity. It measures the actual trading volume of the token on OKX DEX and is the main driver of the ranking.

• Liquidity (r = −0.777, True Signal). It exhibits a threshold effect rather than a linear relationship. Below $500K, it's almost impossible to be in the top 20; below $100K, it's generally ranked after the 30th. Liquidity conveys different information than trading volume; trading volume shows "how much was traded," while liquidity shows "whether the market can absorb it."

• The number of unique addresses (r = −0.750, true signal). The original correlation is very strong, and the partial correlation, after removing liquidity, still shows r = −0.313 (p = 0.034), which is still significant. This measures how many different people are participating in the transaction, serving as a proxy indicator for user dispersion and ecosystem health.

• Active address ratio (r = −0.550, partial signal). The number of unique addresses divided by the total number of cryptocurrency-holding addresses, measuring "how many cryptocurrency holders are actually trading". It has independent informational value, but partially overlaps with trading volume, and its predictive power is weak on its own.

• Liquidity/Holdings (r = −0.512, partial signal). The liquidity depth corresponding to each holding address, measuring the quality of holdings rather than the quantity. Highly correlated with liquidity itself, with limited additional contribution.

• Trading volume/market capitalization (r = −0.407, partial signal). This is the turnover rate, reflecting the trading activity of a token relative to its market capitalization. It has some independent signal, but the sample shows significant fluctuations.

• The transaction volume to address ratio (r = −0.644, pseudo-correlation). The original correlation appears strong, but after removing the influence of transaction volume, the partial correlation is only r = −0.03 (p = 0.85), which is completely insignificant. A high transaction volume ratio is simply a natural consequence of high transaction volume, not an independent signal of penalties for inflated transaction volume.

• Market capitalization (r = −0.751, pseudo-correlation). The original correlation appears to be as strong as trading volume, but the partial correlation, after removing trading volume, is only r = −0.15 (p = 0.31). Tokens with large market capitalization usually also have large trading volumes, and the correlation is entirely explained by trading volume. It is very likely that market capitalization is not even an input variable in the OKX scoring system.

• Number of token-holding addresses (r = −0.148, pseudo-correlation). Direct correlation is extremely weak. RAIN has 170K token holders, ranking 24th, and GAIX has 141K token holders, ranking 41st; these are typical counterexamples. Holding a lot of tokens but not trading them is equivalent to invalid data in this system.

• Price fluctuations (r = −0.019, pseudo-correlation). Linear correlation is almost zero. The impact of price on ranking is not linear, but rather triggers a non-linear penalty during extreme drops (e.g., −52%), while daily positive and negative fluctuations have virtually no effect on ranking.

Figure 2 shows the liquidity situation. The horizontal axis represents the logarithm of liquidity, and the vertical axis represents the ranking (the smaller the number, the higher the ranking). The size of the bubble represents the number of addresses holding the cryptocurrency, and the color represents the price trend (green = rising, red = plummeting, purple = stable).

The most crucial visual information is that the data points generally trend from bottom left to top right, meaning higher liquidity correlates with higher rankings. The $500K mark is a clear dividing line, with almost all of the top 20 tokens to its right. Two large red bubbles in the bottom right corner represent LAB and UB—too liquid to be ranked due to their price crash. The bottom left corner is filled with numerous small bubbles, representing "zombie tokens" with near-zero liquidity and rankings below 30. RAIN is the most noticeable anomaly, with an extremely large bubble (170K holdings) but very low liquidity, placing it at 24th.

Figure 3, as shown above, has a similar structure to Figure 2, but the horizontal axis is replaced with trading volume. The trend and liquidity charts are highly similar, indicating that both influence the ranking in the same direction. The most important information is the two red outliers—LAB ($222M trading volume but ranked 4th) and UB ($79M but ranked 8th)—which stand out significantly on the horizontal axis compared to all other tokens, but their rankings on the vertical axis are not correspondingly higher. This directly illustrates that high trading volume is discounted by the system during price crashes or abnormal TX ratios ; more trading volume is not necessarily better, quality is more important than quantity.

Figure 4 shows the horizontal axis as the number of unique transaction addresses (in K) and the vertical axis as the ranking. The size of the bubbles represents liquidity depth. Colors distinguish three types of tokens: purple for common tokens, green for new tokens (≤10 days old), and red for outliers. The figure shows a loose positive correlation—the more unique addresses, the higher the ranking generally—but this relationship is far from linear and highly discrete. This aligns with the partial correlation coefficient r = −0.31: the number of unique addresses has independent predictive power, but its explanatory power is limited and it is not the sole determinant of ranking.

Most common tokens are concentrated in the range of 0–2K unique addresses, while their rankings range from 1 to 46. This shows that the number of unique addresses alone cannot accurately predict rankings—even with the same number of 1K addresses, tokens with higher liquidity (larger bubbles) rank significantly higher, confirming that liquidity is a necessary condition in addition to the number of unique addresses.

The number of unique addresses for new tokens like AI, OPG, BLEND, PROS, and ASSET is generally not high, but their rankings are in the middle or even higher. This is a direct reflection of the +50% new entry bonus; OKX Boost's bonus mechanism has pulled them up from their expected ranking positions.

OKX is looking for "tokens that are actually being bought and sold," not "tokens that are held by a large number of people." Liquidity is the entry threshold (those with less than ~$500K are basically out of the top 20), trading volume is a signal of activity, and independent addresses remove the interference of wash trading.

2. X Launch Development Status

As of March 3, 2026, a total of 68 X Launch projects have been closed, with an average of 29,949 participants and a peak of 68,900 participants. The highest number of participants was for this project, Avantis.

The top three projects in terms of the number of participants are Avantis, Plasma, and Audiera.

• The precipitous drop in the number of participants: The most obvious pattern is that from November to December 2025, the average number of participants plummeted from 37,000 to 20,000, almost halved, and then remained at around 20,000 until it slightly rebounded in March-April 2026.

• The activity duration is highly standardized: over 78% of projects use a 24-hour activity duration, indicating that the platform has formed a "standard product." However, a few projects still use 48 hours, 72 hours, or even 94 hours.

• The reward value exhibits a clearly right-skewed distribution. The median is approximately $499K, while the mean is a high $901K. A few outliers (Plasma $10.7M, PerpStock $8M, Midnight $4.1M, LINEA $3.8M) significantly inflate the mean. Most projects fall within the $200K–$700K range.

• The number of participants is almost unrelated to USD rewards. A prime example: Avantis attracted 68,900 participants, but the average reward was only $11; while PerpStock and Spacecoin had far fewer participants, yet the average rewards were a whopping $381 and $129 respectively. Participation numbers reflect platform popularity, while reward size reflects project budget and token price performance after launch.

• User engagement experienced a precipitous decline. Average engagement dropped from approximately 45,000 in September 2025 to approximately 17,000 in February 2026, a decrease of 62%, before slightly recovering to approximately 24,000 in April. This decline was not accidental, but closely mirrored the cooling of overall market sentiment, and may also reflect that growth among the early user base had reached saturation.

• Average USD reward per user is the most valuable metric. Most projects offer average returns between $10 and $30. Projects exceeding $50 (PerpStock, Spacecoin, LINEA, Plasma, Midnight, Lombard) are rare exceptions, either due to strong token prices after listing or exceptionally generous initial reward pools.

• The duration of an event has little impact on the number of participants. 24 hours has become the de facto standard, and the few events that lasted more than 48 hours did not show a systematic increase in participation.

The above is a ranking of projects with a total distributed token value exceeding 500k when users claim X Launch airdrop tokens.

3. X Launch data with a more transparent mechanism

The five latest projects launched on X Launch have disclosed their total Boost amount. The average Boost amount per user ranges from 750 to 870. The average reward per user varies depending on the project. This relatively consistent average Boost amount suggests that users' trading on X Launch is becoming increasingly sophisticated, with more "semi-professional" players actively participating.

4. User Cost Analysis

Slippage cost = Actual trade volume × Slippage % × Two round trips + DEX fee + Gas. We calculated the cost under different trading paths, assuming an average Boost volume of $820 per person, and found that the overall conservative trading cost of BSC is around $3.5.

Assuming a slippage rate of 0.5%, the slippage fee would be approximately $9.65, resulting in a total cost of around $13. This means that if the airdrop value is less than $13, it would result in a loss.

The above graph is a break-even point matrix for users under different slippage and different revenue scenarios. In this graph, assuming an average reward value of $20 per person for the five most recent projects, a slippage of 1% will reach the break-even point; a slippage of 1.5% will result in a loss.

In our data, the median average reward for the 68 projects is only about $18, which means that most ordinary users are actually losing money by participating in X Launch if they do not carefully select trading pairs. However, the loss is not in transaction fees, but in the slippage of buying and selling memes without them realizing it.

5. Multi-party game mechanism

• Users are the starting point of the entire chain, with an average of about 24,000 participants per phase. Users conduct actual transactions on the DEX, generating Boost volume that is passed on to the project team, while also paying DEX fees (0.1–0.2%, about $2–3 per person) to the OKX platform, and incurring slippage losses ($10–60 per person). This portion flows to LPs/liquidity pools.
• The project team , as the initiator of the token reward pool, receives users' transaction behavior data and distributes token rewards to users. The size of each period ranges from $123K to $10.7M, with large fluctuations.
• OKX, as the central platform, collects DEX transaction fees and also enjoys the proceeds from token listings. The platform then forwards transaction requests to the blockchain, where chain validators (Solana or EVM network) complete the settlement. Users need to pay gas fees; Solana is almost free, while EVM costs approximately $0.5–$2 per person.
• As market makers for the Meme pool, LPs/liquidity pools passively absorb slippage losses from user transactions and are thus implicit beneficiaries.

Overall, funds flow out from the user side and disperse to OKX, LPs, and chain validators. Project teams, on the other hand, drive user transaction volume through token incentives, forming a closed loop centered on transaction activity.

While some of the value from this segment spills over to limited partners (LPs) on other public chains, most OKX Launches only last about 24 hours, so the spillover effect is minimal. After the X Launch ends, inflating trading volume on the OKX Boost leaderboard becomes OKX's primary source of revenue. For example, on May 3, 2026, the 24-hour trading volume was 371.88 million, and the daily DEX fees were approximately 3.7-7.4 million.