MEV BOT BUILDING: Struggles with Sandwiching Aggregators & Accurate Slippage Calculation (v6 Ethers)

Hi everyone,

I’ve been developing a MEV Sandwich bot using TypeScript (Ethers v6) and running my own execution node. While I’ve managed to get the basic pipeline working—receiving mempool transactions and sending Flashbots bundles—I’ve hit a massive wall regarding complexity and competition. I’m hoping to get some perspective on whether I’m missing a fundamental trick or if the Ethereum "sandwich" game is simply closed to solo devs at this point.

The Technical Hurdle: Aggregator Complexity

The main issue I'm facing is decoding and calculating slippage for transactions routed through complex aggregators (1inch, Uniswap Universal Router, etc.).

Decoding: These calls are deeply encoded. Extracting the true "minimum out" is a nightmare.

Hidden Fees: Aggregators often take their own spread or fees mid-route.

Slippage Math: Because I can't perfectly account for these internal fees, my slippage calculations are slightly off. This causes my bundles to fail or, in the few cases they landed, I made a $0.01 profit but lost $0.02 to unaccounted fees.

The "Jared" Problem (Competition)

I’m constantly being out-competed by top-tier searchers (like the infamous jaredfromsubway.eth). Even when I try to trace logs to understand the state changes, the data volume is overwhelming.

Latency: My node isn't in a prime location (like AWS-East or the Netherlands), so I suspect I'm seeing transactions significantly later than the top bots who are physically peered with the major builders.

Precision: These top bots seem to have "perfect" math for even the most obscure aggregator routes, hitting the slippage limit to the exact cent.

My Questions to the Community:

Decoding Strategies: For those handling aggregators, are you manually writing decoders for every router version, or is there a more generalized approach to simulating the amountOut after all internal fees?

Trace Analysis: When looking at debug_traceCall, how do you efficiently filter the noise to find the actual swap constraints in a sea of nested calls?

Is Mainnet Dead for Solo Searchers? I’ve seen people suggest that Ethereum Mainnet is now a "game of giants" and that new devs should look at AppChains or L2 subnets where there might still be "crumbs" left. Do you agree?

I love the technical challenge, but competing for a $0.01 profit while fighting against bots with better peering and better math is becoming a bit soul-crushing.

Any advice, resources, or reality checks would be greatly appreciated!

PS: All infromation was generated by AI, because explain that issues for me hard :) I hope we could resolve my misunderstanding.

PSS: If you have new idea or method to make money using such tech(I mean not only sandwich), it'll be nice if you could share with me