Dispatch

Most retail traders imagine MEV bots as clever scripts running on cloud servers. The reality is closer to a co-located high-frequency trading operation. Understanding the gap between what you're running and what they're running is essential to understanding why you keep losing.

The 200ms advantage from ShredStream alone is the difference between catching a pool at entry and catching it after the first wave of bots has already pushed the price 15%. You are not competing against other retail traders. You are competing against bare-metal infrastructure with a 200ms head start and private visibility into every transaction you submit.

The Jito stack has five components. The validator client (Jito-Solana, running on ~95% of active stake by 2026). The Block Engine (off-chain auction that receives, simulates, and scores bundles). The Relayer (network layer that introduces a deliberate ~200ms delay to make bundling viable). The Tip Payment Program (on-chain tip receiver). And the Tip Distribution Program (splits tips between validators and stakers).

That relayer delay is the critical detail. The 200ms window is not a bug. It is architectural. It exists to give searchers time to see your transaction, construct a sandwich, and submit the bundle before the slot closes. The infrastructure is literally designed to give the extractor time to extract.

Bundles have hard constraints: maximum 5 transactions, sequential atomic execution in a single slot, no cross-slot spanning, and the tip instruction must be the last transaction. The minimum practical tip for competitive opportunities is 0.01-0.5 SOL. The minimum profit threshold for a viable sandwich is roughly 0.005 SOL. Below that, the gas and tip costs exceed the extraction.

Maybe a dozen serious MEV operations exist on Solana. They have bare-metal servers co-located with validators, custom Rust infrastructure, and private validator relationships. The solo operator success rate is below 5%. This is not a level playing field. It is a stacked deck. The only way to win is to refuse to play their game and route around the infrastructure they control.

Solana produces blocks every 400ms. The first 200ms of each slot is where MEV extraction is decided. Understanding this window is the difference between understanding why you're losing money on every trade and thinking the market is just "unlucky."

Here is what actually happens when you submit a swap on a Solana DEX through a standard wallet. Your transaction enters the public mempool. MEV searchers running Jito's block engine scan that mempool in real-time. When they detect a large swap with slippage tolerance above 1%, they construct a sandwich bundle: three transactions atomically linked. A front-run buy, your original swap, and a back-run sell. The bundle is submitted to the block engine with an attached tip.

The block engine filters invalid bundles and conducts an auction among the remaining ones. The bundle with the highest tip wins priority execution. The validator includes it. The sandwich executes. You receive worse-than-market price on your swap, and the difference between what you should have paid and what you actually paid goes to the searcher, minus the tip to the validator.

You are paying for your own extraction. The tip that the MEV searcher pays to the validator comes from the profit they made by front-running you. The validator profits. The searcher profits. You lose on the spread, the slippage, and the gas. Three layers of extraction on a single trade.

The countermeasure is not higher slippage or faster clicking. The countermeasure is never entering the public mempool. Private RPC routing and Jito tip transactions that bypass public visibility mean the searcher never sees your transaction. They cannot sandwich what they cannot observe. This is not a theory. It is the foundational architecture of ZeroSnipe's execution layer.

When Jito suspended their mempool in March 2024, sandwich attacks dropped to near-zero on bundled transactions. The infrastructure for extraction was always the mempool. Close the mempool, close the extraction. The problem is that most retail traders still route through public channels because their wallets default to it.

Academic researchers recently published a benchmark of 117 confirmed rug pull tokens on Solana, drawn from 68 community-reported incidents. The findings distill into three execution methods. If you trade memecoins on Solana, you should know all three by name.

The numbers are the lesson. Pump-and-dump represents 79% of confirmed rugs. This is not the dramatic "dev pulled liquidity" event that makes CT threads. It is coordinated promotion followed by silent selling. The token doesn't crash. It bleeds. By the time the community notices, the dev wallet is empty and the holders are bagging a slow-decaying asset they'll hold for months hoping for recovery.

Liquidity withdrawal is the classic rug. The developer injects initial capital into a Raydium or PumpSwap pool, receives LP tokens in return, and then quietly retains those tokens. Once the pool reaches sufficient depth (read: once enough retail capital has entered), the developer redeems their LP tokens and drains the pool. The token price doesn't just drop. It goes to zero instantly. There is no exit. The liquidity is gone.

Freeze authority abuse is the most insidious. The SPL token standard on Solana allows the mint authority to grant a freeze authority. This authority can freeze individual token accounts, preventing the holder from transferring or selling. The developer doesn't need to pull liquidity. They simply freeze every wallet except their own, dump their position into the remaining liquidity, and walk away. Victims watch the price collapse in real-time, unable to act.

The validation countermeasure is straightforward. Before entering any position, verify:

mint_authority == null OR revoked
freeze_authority == null OR revoked
LP_tokens == locked OR burned
top_10_holders < 50% of supply

If any of these checks fail, the token is a conditional rug. The developer has not pulled yet, but they have the authority to do so at any time. You are trading on the assumption that someone with the power to rob you will choose not to. That is not a trade. It is a gift.

Legitimate projects renounce both authorities and lock LP tokens publicly. If a project has not done this, it is not "early." It is unvetted. There is a difference.

Every token on pump.fun starts life on a bonding curve. Most traders interact with this curve every day without understanding the mathematics that govern it. This is by design. The curve is the first extraction layer, and understanding it is the first step toward not being extracted by it.

The curve is deterministic. Every buy pushes the price up along a fixed formula. Every sell pushes it back down along the same formula. There is no order book. There is no market maker. There is no negotiation. The curve quotes you a price based solely on cumulative buy volume, and that price is final.

The implication most traders miss: the curve is symmetric. If a token goes from $6,000 market cap to $69,000 (graduation), the price has multiplied roughly 10x. If sellers then exit, the price walks back down the same slope. The last buyers before the decline absorb nearly the entire loss.

Consider a concrete example. You enter at $6,000 market cap with $50. The token catches attention and climbs toward graduation. Your $50 is now worth roughly $500 on paper. You feel like a genius. But the buyer who entered at $60,000 market cap paid roughly 10x your price for the same tokens. When attention leaves and the curve walks back down, that buyer's position goes to near-zero while yours remains marginally profitable.

The curve does not reward skill. It rewards queue position. Being early is not alpha. It is luck in sequencing. The 1% fee extracts on every trade regardless of direction, meaning the platform profits from both the pump and the dump. The curve is not a market. It is a vending machine that raises the price with every sale and refunds along the same schedule.

Graduation changes the game but does not improve the odds. When a token graduates, its liquidity migrates to a DEX pool. The token now trades against paired SOL in an AMM. This is where MEV extraction takes over from bonding curve extraction. The graduation threshold is not a safety certificate. It is the transition from Phase 1 extraction to Phase 2 extraction.

Of every token launched on pump.fun, fewer than 2% reach graduation. Of those that graduate, the majority lose 90%+ of their value within 48 hours. Graduation is survival of the first filter. It is not evidence of long-term viability.

This is the first dispatch. It is not a technical analysis. It is an explanation of why this system exists.

I spent years building trading systems in traditional finance. When I started trading Solana memecoins, I expected to find an inefficient market I could profit from.

What I found was a market that works fine -- just not for retail traders. The bonding curve, the DEX pools, the MEV infrastructure: each layer is optimized to move money from people who don't understand the mechanics to people who do. It's not a conspiracy. It's just incentives, and the incentives point one way.

The tools retail traders use don't help. Your wallet routes through the public mempool. The launchpad shows bonding curve progress but not holder distribution. The DEX aggregator optimizes for price impact but ignores MEV exposure. Everything gives you a sense of control while the extraction happens at a layer you can't see.

I built ZeroSnipe because I was tired of watching good traders lose money to infrastructure they couldn't see. The system isn't magic. It just routes around the extraction instead of through it.

More dispatches coming. Each one will break down a specific part of how the market actually works -- who profits, how, and what you can do about it.