Base Transaction Benchmark: BlockRazor Achieves a 67.75% Inclusion Win Rate

Base transaction submission benchmark: BlockRazor won 813 of 1,200 same-nonce races across 12 runs

Base's Denim upgrade reduced the native block interval from 2 seconds to 200ms, tightening the window for transaction delivery. For teams running arbitrage, liquidation, and market-making strategies on Base, timely inclusion is an important measure of a submission service. See the official native block specification and v1.5.1 release notes for upgrade details.

To show how BlockRazor performs across different times of day, we ran same-nonce transaction submission tests from a single client in Virginia, USA. This article focuses on inclusion results for BlockRazor and Base's official Sequencer. The test comprised 12 runs starting every 2 hours, with 100 races per run.

Across all 1,200 races, BlockRazor Virginia won 813, achieving a 67.75% inclusion win rate. Base's official Sequencer won 387 (32.25%). BlockRazor led in 9 of 12 runs, securing more than two-thirds of the winning inclusions.

Benchmark Setup: 12 Runs Every 2 Hours

This article presents results for the following transaction submission paths on Base Mainnet (Chain ID 8453):

Submission Path	Endpoint Tested
BlockRazor Virginia	`https://virginia-base.blockrazor.io:10101/eth_sendRawTransaction`
Base Official Sequencer	`https://mainnet-sequencer.base.org`

The test took place on November 7–8, 2026. Each run included 100 races, with a 2-second cooldown configured between races. Runs started at 2-hour intervals, with 22 hours between the first and last starts.

Test size: 12 runs and 1,200 races.
Winner determination: 1,200 successful transaction receipts, checked through Base's official public node at https://mainnet.base.org.
Round-start block range: 54,361,773–54,758,811.
Winning inclusion block range: 54,361,775–54,758,813.

Microsecond Submission Timing: Comparing Request Start Order

To verify synchronized submission, we compared the client-side request start times for BlockRazor and the official Sequencer in each of the 1,200 races. We measured the absolute difference between their request start times:

Submission Timing Metric	BlockRazor–Sequencer Request Start Difference
P50 (median)	4 μs
P90	30 μs
Maximum	67 μs

The submission order did not consistently put BlockRazor first. Among the 1,083 races with distinguishable request start times, BlockRazor started earlier 552 times, and the official Sequencer started earlier 531 times. Another 117 races had simultaneous starts at microsecond precision.

In the 531 races where the official Sequencer started its request before BlockRazor, BlockRazor still won 371 times. These results demonstrate winning inclusion even when the Sequencer's request started earlier; BlockRazor's wins did not require a fixed first-start advantage.

Methodology: Same Nonce, Matched Fees, Onchain Winners

In each race, the tested routes submitted transactions using the same sender and nonce. A nonce can be consumed only once on the canonical chain. We matched the transaction hash in the successful receipt to its submission route to determine the winner.

To help users compare transaction delivery under matched fee conditions, we applied consistent parameters for the compared transactions:

Matched transactions and fees: each race used the same sender, recipient, nonce, chain ID, value, empty calldata, and fee parameters. All transactions were EIP-1559 Type 2 zero-value transfers, with maxPriorityFeePerGas fixed at 1,000,000 wei (0.001 gwei).
Identifiable transaction hashes: gas limits distinguished the transactions: 30,000 for BlockRazor and 30,002 for the Sequencer. Every winning receipt used 21,000 gas and returned status=1.
Concurrent submission: the client submitted transactions concurrently, allowing them to compete for inclusion with the same nonce and fees.

Overall Results: BlockRazor Won 813 Races at 67.75%

Submission Path	Races Won	Inclusion Win Rate
BlockRazor Virginia	813	67.75%
Base Official Sequencer	387	32.25%
Total	1,200	100.00%

With the same nonce and matched fees, BlockRazor won more than twice as many races as the official Sequencer. For teams focused on execution efficiency, this result demonstrates BlockRazor's transaction delivery competitiveness in this Base Mainnet test.

All 12 Runs: BlockRazor Led in 9

Each run contains 100 races, so each percentage also equals the number of races won. T+00h marks the first run's start; subsequent runs are ordered by their start interval relative to that point.

Run	Time Since First Start	BlockRazor	Sequencer
01	T+00h	72%	28%
02	T+02h	91%	9%
03	T+04h	48%	52%
04	T+06h	100%	0%
05	T+08h	47%	53%
06	T+10h	83%	17%
07	T+12h	66%	34%
08	T+14h	62%	38%
09	T+16h	57%	43%
10	T+18h	31%	69%
11	T+20h	68%	32%
12	T+22h	88%	12%
Total / pooled rate	1,200 races	813 / 67.75%	387 / 32.25%

BlockRazor led in 9 runs, reaching 100% in run 4, 91% in run 2, and 88% in run 12. The official Sequencer won more races in runs 3, 5, and 10. All 12 runs contribute to BlockRazor's 67.75% overall inclusion win rate.

Evaluate Base Transaction Delivery Through Onchain Results

For strategies that depend on timely inclusion, a submission service's value ultimately needs to appear in onchain execution results. Across 12 test windows and 1,200 races, BlockRazor delivered 813 wins, a 67.75% overall win rate, and a lead in 9 test windows under matched priority fees.

We welcome teams running arbitrage, liquidation, and market-making strategies on Base to try BlockRazor with their own transaction workloads and fee policies, and evaluate execution in their deployment environments.

Read the BlockRazor Base Transaction Submission Configuration Guide.
Join the BlockRazor Discord community for technical support and test access.