Why are so many transactions unconfirmed on Testnet3?

Just noticed something interesting for those using testnet3. Tons of transactions show up on this mempool site and they sit unconfirmed for days. It’s weird since some blocks are only half full, and those transactions could easily fit in. But why are so many blocks empty? It’s like people send their transactions once then just close their clients.

You’d be shocked at how few mainnet users actually run their own nodes. Seriously, it’s a small number. And don’t forget about the miners. If they’re only making money through transaction fees, they really should allow more transactions in their mempool and keep them longer.

Honestly, these issues can pop up on any chain that's just copying Bitcoin Core. Even if you’re a casual user, you really should run a node to rebroadcast your own transactions. That’s why we have testnet4. Every 2016 blocks, the difficulty gets a reset. If you mine at the right time, the code thinks it’s been two weeks of network difficulty.

So, on testnet3, if no block is found for 20 minutes, it’s possible to mine a long chain of easy blocks and broadcast them right away. But testnet4 won’t let that chain go further than 2 hours ahead, and each block needs a time gap of 20 minutes and 1 second. Am I right?

Yep, that’s the time-warp attack! That’s why BIP-94 was created. There’s a two-hour rule across all networks. Without BIP-94, you can fake timestamps, claiming something took two weeks when it didn’t. If you adjust the difficulty mid-way, you can end up mining six blocks per second, which is just insane.

Running a node doesn’t guarantee your transaction gets mined. Once a transaction hits the network, it goes into other nodes’ mempools. When a new miner connects, they grab the mempool data from others. So if someone goes offline, their transactions can still get picked up.

For real? I saw the transaction count jump from 100k to 400k, but my node didn’t see any of that. Looks like someone is pushing transactions to their block explorer without sharing them across the network.

Mempool.space probably has its own glitches. In testnet4, there’s an address showing a negative balance of over 20k bitcoin. Must be a backend miscalculation when the chain reorganizes. That happens when a miner reorganizes while a transaction spending from their address is also being reorganized.