Trading Strategies

This section covers the various execution strategies available in arb-assist, including spam mode, Jito bundles, and other advanced trading configurations.

Strategy Levels Overview

arb-assist supports multiple trading strategies that activate based on market conditions:

# Each strategy maps to filter threshold levels
spam_levels = [...]      # Standard fast execution
jito_levels = [...]      # MEV-protected bundles
fast_levels = [...]      # Premium fast lanes
nextblock_levels = [...]  # Next block inclusion

Each level corresponds to a filter threshold tier, allowing different strategies for different market conditions.

Spam Strategy

Configuration

spam_levels = [
  {
    filter_level = 0,                    # Maps to filter_thresholds[0]
    bundle_groups = [1],                 # Which mint groups to use
    process_delay = 400,                 # Delay between attempts (ms)
    min_cu_percentile = 25,              # Minimum fee percentile
    max_cu_percentile = 50,              # Maximum fee percentile
    min_cu_price = 1_000,               # Minimum compute unit price
    max_cu_price = 10_000,              # Maximum compute unit price
    tx_count = 1,                       # Transactions per burst
    fee_strategy = "Random"             # Fee selection strategy
  },
  {
    filter_level = 1,                    # Higher tier
    bundle_groups = [1, 2],             # More groups
    process_delay = 300,                # Faster retry
    min_cu_percentile = 50,
    max_cu_percentile = 75,
    min_cu_price = 10_000,
    max_cu_price = 50_000,
    tx_count = 3,                       # More aggressive
    fee_strategy = "Linear"
  }
]

Fee Strategies

Random

fee_strategy = "Random"

• Randomly selects fee within range

• Good for avoiding detection

• Unpredictable cost

Linear

fee_strategy = "Linear"

• Gradually increases fees

• Predictable escalation

• Good for testing limits

Exponential

fee_strategy = "Exponential"

• Rapidly increasing fees

• For highly competitive situations

• Can get expensive quickly

Priority Fee Sources

When helius_key is configured:

helius_key = "your-api-key"

• Uses real-time network fee data

• Percentiles based on recent blocks

• More accurate pricing

Without Helius:

• Learns from parsed transactions

• Less accurate but functional

• May overpay or underpay

Jito Bundle Strategy

Basic Configuration

jito_levels = [
  {
    filter_level = 0,
    bundle_groups = [1, 2],
    process_delay = 400,
    min_tip = 1_000,                    # Minimum Jito tip
    max_tip = 10_000,                   # Maximum Jito tip
    min_tip_percentile = 25,            # Dynamic tip percentile
    max_tip_percentile = 50,
    tx_count = 1,
    tip_strategy = "Random",
    sending_strategy = "AllAtOnce",     # Bundle sending method
    no_failure_mode = true              # Ensure bundle success
  }
]

Jito-Specific Settings

no_failure_mode

no_failure_mode = true   # All transactions must succeed
no_failure_mode = false  # Allow partial failures

• true: Safer but consumes tip always

• false: Risk of wasted tips

sending_strategy

sending_strategy = "AllAtOnce"  # Send complete bundle
sending_strategy = "OneByOne"   # Send individually

Dynamic Jito Tips

dynamic_jito_tip_mode = "parsed"     # Learn from blockchain
dynamic_jito_tip_mode = "tipstream"  # Use tip stream API
dynamic_jito_tip_mode = "none"       # Fixed tips only

Parsed Mode

• Analyzes successful bundles

• Learns optimal tip amounts

• Self-adjusting

Tipstream Mode

• Real-time tip floor data

• Most accurate pricing

• Requires API access

Jito Configuration (SMB)

[smb.jito_config]
ips = ["192.168.1.0/24"]              # Allowed IP ranges
min_profit = 5_000                    # Minimum profit to use Jito
use_min_profit = true                 # Enforce minimum
use_separate_tip_account = false      # Tip account management

Jito Configuration (NotArb)

[notarb]
jito_targets = [
  "https://ny.mainnet.block-engine.jito.wtf",
  "https://tokyo.mainnet.block-engine.jito.wtf",
  "https://amsterdam.mainnet.block-engine.jito.wtf",
]
max_bundle_transactions = 5           # Bundle size limit

Advanced Trading Strategies

Fast Lane Services

fast_levels = [
  {
    filter_level = 2,                   # High-value opportunities
    bundle_groups = [1],
    process_delay = 200,
    min_tip = 1_000_000,               # Higher tips
    max_tip = 2_000_000,
    tip_strategy = "Linear",
    no_failure_mode = false,
    min_cu_price = 50_000,
    max_cu_price = 100_000
  }
]

[smb.fast_config]
min_profit = 50_000
use_min_profit = true
auth_value = "your-fast-lane-auth"

NextBlock Strategy

nextblock_levels = [
  {
    filter_level = 2,
    bundle_groups = [1],
    min_tip = 2_000_000,
    max_tip = 5_000_000,
    tip_strategy = "Exponential"
  }
]

[smb.nextblock_config]
min_profit = 100_000
use_min_profit = true
auth_value = "your-nextblock-auth"

Multi-Strategy Coordination

Cascading Strategies

Configure multiple strategies for different market conditions:

# Tier 0: Normal market - Use spam
spam_levels = [
  { filter_level = 0, bundle_groups = [1, 2], ... }
]

# Tier 1: Competitive - Add Jito
jito_levels = [
  { filter_level = 1, bundle_groups = [3, 4], ... }
]

# Tier 2: High value - Use premium services
fast_levels = [
  { filter_level = 2, bundle_groups = [5], ... }
]

Bundle Group Strategy

Assign different groups to different strategies:

mints_to_arb = [2, 2, 3, 3, 4]  # 5 groups

# Groups 1-2: Spam strategy (low value)
# Groups 3-4: Jito bundles (medium value)
# Group 5: Fast lane (high value)

Timing Configuration

Process Delays

process_delay = 400  # Standard delay (ms)

Recommendations:

• 200-400ms: Aggressive, may fail more

• 400-600ms: Balanced approach

• 600-1000ms: Conservative, higher success

Update Intervals

update_interval = 10000  # Config update frequency (ms)
run_interval = 60000     # Minimum time between trades

Balance between:

• Fresh data (lower interval)

• System resources (higher interval)

• Market volatility (adjust accordingly)

Risk Management

Minimum Profit Requirements

Always set minimum profit thresholds:

[smb.jito_config]
min_profit = 10_000      # Only use Jito if profitable
use_min_profit = true

Transaction Limits

Control exposure with transaction counts:

tx_count = 1  # Conservative
tx_count = 3  # Moderate
tx_count = 5  # Aggressive

Compute Unit Limits

Prevent excessive fees:

max_cu_limit = 400_000   # Standard
max_cu_limit = 600_000   # Complex routes
max_cu_limit = 800_000   # Maximum allowed

Strategy Selection Guide

For New Users

Start conservative:

# Single strategy, moderate fees
spam_levels = [{
  filter_level = 0,
  bundle_groups = [1],
  min_cu_price = 5_000,
  max_cu_price = 20_000,
  tx_count = 1,
  fee_strategy = "Random"
}]

For Competitive Markets

Use Jito with dynamic tips:

dynamic_jito_tip_mode = "parsed"

jito_levels = [{
  filter_level = 0,
  bundle_groups = [1, 2],
  min_tip_percentile = 50,
  max_tip_percentile = 90,
  no_failure_mode = true
}]

For High-Value Opportunities

Combine multiple strategies:

# Use everything available
spam_levels = [{ filter_level = 0, ... }]
jito_levels = [{ filter_level = 1, ... }]
fast_levels = [{ filter_level = 2, ... }]

Performance Optimization

Fee Optimization

Monitor and adjust:

1. Track success rates by fee level

2. Find optimal fee ranges

3. Adjust percentiles based on results

4. Use Helius for better estimates

Bundle Optimization

For Jito bundles:

1. Group similar value mints

2. Keep bundles small (1-3 txs)

3. Monitor tip efficiency

4. Adjust based on competition

Latency Optimization

Reduce delays:

1. Colocate with RPC nodes

2. Use multiple sending RPCs

3. Optimize process_delay

4. Monitor network conditions

Common Issues

High Failure Rate

Symptoms:

• Many failed transactions

• Wasted fees

• Low profitability

Solutions:

• Increase process_delay

• Raise priority fees

• Switch to Jito bundles

• Check RPC performance

Overpaying Fees

Symptoms:

• High gas costs

• Low ROI

• Profitable trades become losses

Solutions:

• Lower fee percentiles

• Use dynamic fee estimation

• Set tighter fee ranges

• Monitor competitor fees

Missing Opportunities

Symptoms:

• Seeing arbs on-chain you didn't catch

• Low transaction count

• Competitors winning

Solutions:

• Reduce process_delay

• Increase fee ranges

• Add more strategies

• Use faster RPCs