Trading Impact Scalar Params
| Market | Current Impact Scalar | Rec Impact Scalar | Change |
|---|---|---|---|
| SOL | 1,250,000,000 | 1,360,000,000 | 8.8% |
| ETH | 5,000,000,000 | 3,500,000,000 | -30% |
| BTC | 8,000,000,000 | 12,000,000,000 | 50% |
Executive Summary: Trading Impact Scalar Analysis
Key Findings
1. SOL:
- Current scalar: 1.25B; Recommended: 1.36B (8.8% change)
- Avg trade: $16,495; Median: $2,550; 99.9th percentile: $1M
2. BTC:
- Current scalar: 8B; Recommended: 12B (50% change)
- Avg trade: $21,810; Median: $2,794; 99.9th percentile: $2M
3. ETH:
- Current scalar: 5B; Recommended: 3.5B (-30% change)
- Avg trade: $24,920; Median: $2,791; 99.9th percentile: $2M
Optimization Models
- Univariate:
- Optimizes impact scalar alone
- SOL: 1.36B, BTC: 12B, ETH: 3.5B
- Bivariate:
- Optimizes impact scalar and base fee simultaneously
- SOL: 1.23B scalar, 0.053% base fee
- BTC: 6.47B scalar, 0.049% base fee
- ETH: 3.13B scalar, 0.052% base fee
Liquidity Metrics
- Order book depth stable over time for all assets
- BTC: >$35M depth beyond ±20 bps
- ETH: ~$15M depth at ±100 bps
Recommendations
- Implement optimized scalars: SOL (1.36B), BTC (12B), ETH (3.5B)
- Consider bivariate model results for further optimization
- Tailor fee structures to accommodate wide trade size range
- Monitor impacts, leveraging consistent order book conditions
Supporting Analysis - SOL Trading Impact Scalar
Fee Optimization:
The current impact scalar is set at 1.25B, while the recommended scalar suggests increasing it to 1.36B. This small adjustment is expected to deliver an 8.8% improvement in fee efficiency. The objective function graph shows a clear trend—initially, the function decreases as the impact scalar increases, hitting its lowest point around the recommended scalar before rising again.
Trade Distribution:
Trade sizes vary widely, with an average size of $16,495 but a median of $2,550, indicating most trades are on the smaller side. The standard deviation of $75,473 points to a few very large trades skewing the average. These outliers are further reflected in the 99th percentile, where trades reach up to $246,742, and a 99.9th percentile with trade sizes hitting $1M. Optimizing fees for smaller, more frequent trades is essential as they make up the bulk of activity. However, fee structures must still account for the higher impact of larger orders to ensure a balanced approach.
Liquidity Analysis:
The centralized order book reveals a typical liquidity pattern, with the deepest liquidity further from the mid-market price. Therefore, adjustments should consider this pattern—particularly for larger trades that might push deeper into the order book, incurring higher impact costs.
Order Book Depth Stability:
The order book depth remains consistent over time, showing stability in liquidity across various basis point levels. This stability implies that any changes to the impact scalar would yield predictable results, making it a straightforward process to implement the proposed optimizations.
Bivariate Optimization:
In the bivariate analysis, both the base fee and impact scalar are optimized simultaneously, with the best combination being a base fee of 0.053% and an impact scalar of 1.23B. This optimization yields an objective function result of 229,163, indicating further improvements compared to the univariate model. Reducing both parameters in tandem results in better overall fee efficiency, particularly for SOL trades.
BTC Price Impact Fee Optimization Analysis
The following analysis provides insights into BTC price impact fee optimization using the displayed data and recommendations.
Fee Optimization:
The current impact scalar for BTC is set at 8,000,000,000, with a recommended increase to 12,000,000,000, representing a significant shift to optimize performance. This adjustment results in an objective function improvement of 50%. The graph shows a clear trend where the objective function steadily declines until it reaches the optimal impact scalar of 12B, after which further increases provide diminishing returns.
This indicates that increasing the impact scalar as suggested is crucial for minimizing price impact on BTC trades.
Trade Distribution:
BTC trade sizes are highly variable. The average trade size is $21,810, but the median trade size is $2,794, indicating that most trades are smaller, though a few large trades skew the average upwards. The standard deviation is notably high at $103,844, confirming the presence of a few large trades that increase the overall variability.
The distribution shows that the 25th percentile trade size is $990, and the 75th percentile is $10,183, meaning most traders engage in relatively small-sized orders. The 99th percentile trade size jumps to $332,304, while the 99.9th percentile reaches $2M. These large outliers suggest that while the majority of trades are small, occasional large orders contribute significantly to the market’s overall activity. Optimizing fees should focus on both ends of this distribution to ensure smaller trades are efficient while preventing large orders from incurring substantial impact.
Liquidity and Order Book Analysis:
The order book depth for BTC follows the standard “V” shape. Liquidity increases significantly at the extremes, with depth exceeding $35M beyond +/- 20 bps. This pattern indicates that mid-range trades are likely to face more price impact unless liquidity is improved.
Given this liquidity profile, adjusting the fee structure to offset the price impact for trades near the mid-point can be critical, especially for large orders that might push through several bps of the order book.
Order Book Depth Over Time:
The order book depth over time shows stability across all basis points, suggesting that liquidity conditions have been relatively constant over the observation period. There is no significant volatility in depth, meaning that the impact scalar adjustments will likely have consistent effects, and any fee optimization will be effective without concerns about fluctuating liquidity.
Bivariate Optimization:
In the bivariate optimization, both impact scalar and base fees are optimized together. The current impact scalar is set at 8B, with the recommended scalar lowered to 6.47B and the base fee reduced slightly from 0.06% to 0.049%. This combination results in an objective function score of 887, significantly improving trade efficiency.
This model suggests that lowering both the impact scalar and base fees results in a more optimal outcome for BTC trades compared to univariate optimization.
For BTC trades on Jupiter, the recommended strategy involves increasing the impact scalar to 12B to achieve a 50% improvement in minimizing price impact. However, a bivariate approach also highlights that reducing the base fee in tandem with a modest impact scalar reduction to 6.47B could yield the best results overall. As BTC trades tend to include many smaller orders with the occasional large one, fee structures should target this mix to optimize both ends of the distribution. The order book remains stable, so these adjustments will likely be consistently effective.
ETH Price Impact Fee Optimization Analysis
The following is an analysis of the price impact fee optimization for ETH trades based on the displayed data and recommendations.
Fee Optimization:
The current impact scalar is set at 5,000,000,000, with a recommended reduction to 3,500,000,000. This adjustment results in a 30% improvement in the objective function. The graph shows the objective function decreases significantly up to around 3.5B, after which it starts to increase again. This clearly suggests that lowering the impact scalar to the recommended value will reduce price impact, making trades more efficient.
Trade Distribution:
ETH trades show a significant spread, with an average trade size of $24,920 and a median of $2,791, indicating most trades are relatively small. The standard deviation is high at $132,739, confirming that a few very large trades skew the average higher.
The 25th percentile trade size is $1,000, while the 75th percentile is $9,797, meaning a majority of trades are clustered in the lower range. At the upper end, the 99th percentile trade size is $467,440, and the 99.9th percentile reaches $2M. This distribution suggests that while most ETH trades are small, the occasional large orders play a significant role in overall market activity. Optimizing fees should target both smaller trades, which dominate the volume, and the occasional large orders that have a disproportionate impact on liquidity.
Liquidity and Order Book Analysis:
The order book depth for ETH follows the familiar liquidity trough pattern, where the depth is lowest at 0 bps impact and increases sharply as you move further from the mid-market price. Depth peaks at around $15M at the extremes of +/- 100 bps, meaning that trades closer to the mid-point are more likely to face higher price impacts.
This liquidity profile suggests that optimizing fees and impact scalars should focus on minimizing impact for trades near the mid-point, especially for large orders that could push through several levels of the order book.
Order Book Depth Over Time:
The order book depth over time appears relatively stable across various bps levels, indicating no significant fluctuations in liquidity. This consistency implies that any changes to the impact scalar or fee structure will likely provide predictable improvements, as there is little volatility in the liquidity environment.
Bivariate Optimization:
In the bivariate optimization, both the impact scalar and base fee are optimized together. The current impact scalar is 5B, and the recommended value is reduced to 3.13B, alongside a reduction in the base fee from 0.06% to 0.052%. The result is an objective function score of 10,736, which represents a further improvement in trading efficiency.
This combined optimization suggests that reducing both the base fee and impact scalar simultaneously would yield the best results for ETH trades, especially given the trade distribution and liquidity characteristics of this market.
For ETH trades, reducing the impact scalar from 5B to 3.5B significantly improves trade efficiency, with a 30% improvement in the objective function. The trade distribution skews towards smaller trades, but occasional large orders necessitate a balanced approach in fee optimization. The order book remains stable, ensuring that these adjustments will likely be consistently effective across different market conditions. Additionally, the bivariate optimization, which includes a reduction in both the base fee and impact scalar, suggests an even more refined strategy for minimizing price impact on ETH trades.
Conclusion
Gauntlet recommends a two-phase implementation approach:
- Initially, the impact scalars will be implemented based on the Univariate optimization model: SOL (1.36B), BTC (12B), and ETH (3.5B).
- After gathering performance data from this initial implementation, revisit the fee structure using the Bivariate optimization model. This second phase would involve adjusting both the impact scalars and trading fees simultaneously to potentially achieve further improvements in trading efficiency.
