The casino has better odds

Context first, so you know this isn’t a mood: for a bit over a year I ran my own indexer. I collected and aggregated Solana memecoin data — trades, transfers, launches, migrations — and before I ever placed a real trade, I did nothing but analytics and backtests on that data. What follows is a subjective opinion, but it’s the kind of subjective opinion that comes with a database attached.

It’s also the long version of rules 7 and 10 from the rules list. Those were the conclusions. This is the invoice.

Start with the risks that take 100%, instantly

Memecoin risk isn’t one number, it’s a hierarchy — and the top of the hierarchy is the class of risks that can take your entire position in a single transaction. Before you think about strategy, entries, or anything else, these are binary filters:

• Don’t touch coins with reversed pairs. A pool created with the token order flipped relative to convention. An honest launch has no reason to do this; it exists to confuse tooling and the people relying on it.
• Don’t touch coins whose pool isn’t locked or burned. If the deployer still controls the liquidity, they can pull it whenever they like — and a price with no liquidity behind it is just a number on a screen. One transaction, and your position is worth exactly nothing.

Only after those come the risks people usually talk about: bundles, holder distribution, and so on.

The trades you can’t see

Here’s the part that took me longest to internalize: far from everything that matters happens on the exchange.

A standard play looks like this. A set of wallets — no shared funders, no linking transactions, nothing your clustering heuristics will catch — steadily buys a coin. On every chart and dashboard this reads as a healthy market: organic demand, growing holders, a coin going up. Then, in one moment, every one of those wallets dumps its tokens onto a master wallet, which now holds the majority of supply — and sells everything. The coin goes to zero.

Think you’ll be the one who spots it and exits in time? Good luck. To catch that move as it happens, you can’t just watch the DEX — you have to watch every transfer of every participant. And if they route the tokens through a couple of disposable intermediate wallets instead of sending directly, you’re now tracking hundreds of millions of transactions a day, if not billions. Millisecond reactions to any of them, live risk recalculation across every position you hold — that takes serious expertise, serious infrastructure, and ideally your own validators, or access to someone’s.

Which brings us, smoothly, to the territory where playing consistently in the green is reserved for teams that already have a couple hundred million dollars on the books, a floor of employees, and connections.

The people who own the pipes

Now add the validator operators to the picture. Some of them:

1. Sell a stream of transactions before they land in a block — other players literally pay to see your order before it exists on-chain.
2. Run a custom, rewritten client that lets them reorder transactions and insert their own before and after yours — the classic sandwich: they buy, you buy the price they just pushed up, they sell into your fill.

You’re not trading against the market. You’re trading against people who operate the market’s plumbing.

Leaderboards, and why copytrading is dead on arrival

Let’s talk about the PnL leaderboards, because everyone navigates by them. People share wallets with each other, pull up the PnL on gmgn or wherever, and go “let’s copytrade this guy.” And every time, mysteriously, the copytrade doesn’t work and the money burns down. How come? For a start: you cannot copytrade a wallet without dissecting its statistics first. But it’s worse than that.

I’ll say it with 99% confidence: copytrading Solana memecoins does not work. There are more big players than you think, their infrastructure is beyond what you imagine, and they run on everything that works. Any window of profit gets computed away almost instantly.

I’m sure of this because it happened to my own strategies — the ones that showed even a hint of potential. One example. I was testing coordinated launch strategies, everything timed by block: each wallet enters at its assigned block, each wallet runs its own scripted strategy — timer, take-profit, stop-loss. To keep real users out of the data, I tested on deliberately worthless coins: no image, “test” for a name, broken metadata. By the second launch, two bots were playing against me. By the fourth, ten bots knew my launch pattern and were trading against it. I ran all ten tests inside ten minutes — this was happening fast.

No human writes code at that speed, and no human does manual analysis at that speed. Either those bots run on some seriously smart dynamic model, or something smarter is feeding them — I don’t know. But the same thing happened to every strategy of mine that showed the slightest edge. Do you think it’s hard for these systems to find wallets that consistently trade in profit and copy them? They’ll do it in seconds. Long before you’ve opened the wallet’s page.

It’s a shame, honestly, because the idea is so seductively simple: find someone who wins and do what they do. Alas. And the numbers agree: not a single one of my backtests — with all fees modeled correctly, priority fees and slippage included — showed a single copytrading result that survived past statistical noise over any real distance. Not for one wallet in my entire indexer.

The verdict

So here’s my subjective conclusion: the Solana memecoin market, at least in its current state, is a con. Go to a casino, buy a lottery ticket, put money on a football match. Your odds will be posted honestly, and they will be better.

Two caveats, so this stays fair:

• This applies to coins below a certain bar of capitalization and liquidity. The billion-cap coins — PEPE, BRETT and friends — are a different game, one that looks much more like standard financial markets, where BTC and SOL price action drives most of the movement. Even there, in theory you could find someone worth copying — but the windows are small, HFT decides everything, the bets are bigger, and the hit rate is far below what pre-migration coins pretend to offer.
• The data backs the pessimism. A year of aggregated launches, and the migration statistics are grim — and they get grimmer once you exclude the scam launches with 40–50% bundles, because a depressing share of what does migrate was never a market in the first place.

Knowledge, paid for — this chapter mostly in time and compute. Take it for free.