How to fix rejected shares

Rejected is not the same as stale

A STALE share was valid work that arrived too late: the network moved to a new block while your share was in flight. A REJECTED share failed validation itself: wrong algorithm, malformed work, below the assigned difficulty, or a duplicate. A small stale percentage is normal on any pool; rejected shares in any quantity mean something in your setup is wrong, and they earn nothing.

The checklist, in order of likelihood

• Algorithm: confirm the miner is started with the coin's exact algorithm. A wrong algo produces hashes the pool cannot accept, and often 100% rejects.
• Miner version: update to the latest miner release. After a network upgrade an old miner keeps producing work for the old rules and every share fails.
• Overclock: back your GPU overclock off to stock and test. An unstable core or memory clock produces work that fails verification only sometimes, which is the classic "5% rejected" signature.
• Duplicate rigs: two rigs accidentally configured with the same worker name and connection can double-submit; give every rig its own worker name.
• Connection: high packet loss or a flaky proxy can corrupt or re-send submissions; test the closest region and a wired connection.

Read what the pool says

The reject reply carries a reason: low difficulty, job not found, duplicate, or malformed. Your miner prints it in its console and log. That one line usually names the fix; do not troubleshoot blind when the reason is printed in front of you.

Each reply is written out in what the pool told your miner, including which ones mean something at your end and which mean nothing is wrong.

When only one rig rejects

If your other rigs are clean, the problem is that machine, not the pool: swap its overclock profile, driver version and riser cables into suspicion in that order. If ALL rigs reject at once, re-check the miner version and algorithm first, then the pool's status page and announcements.