Most staking guides tell you to pick a validator with a low commission rate and move on. That's not wrong, exactly — commission matters — but it's treating one variable as if it's the whole picture.
Choosing a validator is really a risk evaluation exercise. You're deciding which operator's infrastructure and behavior is acceptable given your situation. Commission is the easiest variable to assess. The harder parts are reliability, skin in the game, and — on some networks — the possibility of having your stake penalized for someone else's mistakes.
Different protocols also weight these decisions differently. On Ethereum, validator selection affects both your yield and the network's client diversity. On Cosmos, a bad validator can cost you during the unbonding window. On Solana, it's mostly about performance and uptime. The criteria overlap, but what matters most varies by network.
What Validators Actually Do
Validators are network participants that process and attest to transactions in proof-of-stake systems. In return for that work, they earn block rewards and transaction fees — a portion of which they pass on to delegators.
The economics are simple: you delegate stake to a validator, they do the computational work, and you split the proceeds according to their commission rate. If they perform well, you earn roughly the network-average yield. If they perform badly — or get caught behaving maliciously — you risk slashing.
The Variables That Actually Matter
Commission rate is the percentage of staking rewards the validator keeps before distributing the rest to delegators. A 5% commission on a 4% APY network means your effective return is roughly 3.8%. A 10% commission drops it to about 3.6%. The difference is real but smaller than most people expect — which means obsessing over commission while ignoring reliability is a bad trade.
Uptime is probably more consequential than commission. A validator that's offline can't attest to blocks, which means your delegation earns nothing during that period. On Ethereum, offline validators also accrue small inactivity penalties. Consistent downtime compounds into meaningful yield drag over months.
Self-stake (also called self-bond on some protocols) is how much of their own capital the validator has staked alongside delegators. More skin in the game generally signals alignment. A validator running on minimal self-stake has limited personal downside if they underperform or make operational mistakes. This isn't a hard filter — some excellent operators run lean — but it's worth factoring in.
Slashing history is a hard filter. Slashing is the automatic, irreversible penalty that happens when a validator double-signs or behaves in ways the protocol treats as malicious. Critically, on many protocols the penalty falls on delegated stake, not just the validator's own funds. Any validator with slashing events in their history needs a credible public post-mortem. If one doesn't exist, that's your answer.
Protocol-Specific Considerations
A few things worth knowing that don't apply universally:
Ethereum: client diversity and MEV practices. Ethereum has multiple validator client implementations — Prysm, Lighthouse, Teku, Nimbus, Lodestar. A supermajority of validators running a single client creates correlated failure risk for the whole network. Some stakers choose to delegate to operators running minority clients as a small contribution to network health. Separately, most validators now use MEV-boost infrastructure to capture additional revenue from transaction ordering. Some operators share this MEV revenue with delegators; others keep it. Worth checking the policy if you're evaluating individual validators rather than a liquid staking protocol.
Cosmos chains: slashing during unbonding. This one catches people off guard. On Cosmos-based chains (ATOM, OSMO, etc.), slashing applies during the 21-day unbonding window. If you've initiated undelegation and your validator gets penalized before the unbonding period expires, your tokens are still at risk. This makes validator quality more consequential than on networks where your exposure ends the moment you signal your exit.
Solana: uptime is everything. Solana's validator economics are heavily performance-weighted. Commission rates vary widely, but the key variable is consistently achieving block production targets. Solana's history of network outages has also concentrated attention on which validators recover quickly and maintain high participation rates across the full cycle.
What You Can Actually Observe
Most of this information is public. Block explorers and staking dashboards — Rated Network for Ethereum, Mintscan for Cosmos, Stakewiz for Solana — publish validator metrics including historical uptime, effectiveness scores, commission rate history, self-stake amounts, and slashing events.
What's harder to observe in advance: operator infrastructure choices — how geographically distributed their nodes are, which client software they run, whether they have failover systems. Larger, established operators often publish this information. It's worth looking for.
One thing to watch: validators who set a low initial commission rate and then raise it unilaterally later. Some protocols allow this with minimal notice. If you're delegating for the long term, check whether there's a history of rate stability or announced increases rather than surprises.
What Would Change This Picture
Distributed validator technology (DVT) — protocols like Obol and SSV — allows a single validator to be operated across multiple node operators, reducing single-point-of-failure risk. If DVT becomes mainstream, the relevant unit of evaluation shifts from a single validator to an operator cluster. The metrics would change.
Restaking platforms have introduced another variable: some validators opt into restaking, meaning delegated stake is also being used to secure other protocols. Whether the additional yield offsets the additional risk depends on specifics you'd need to evaluate per operator. Worth knowing exists, even if it doesn't affect most delegators today.
The core selection framework — commission, uptime, self-stake, slashing history — remains stable. The ecosystem around it is adding complexity.
Signals to Watch
Confirmation that a validator's selection criteria are sound: consistent uptime above 98% across multiple reporting periods, a commission rate history with stable or announced changes (not unilateral mid-stake increases), self-stake amounts available on-chain, and a clean slashing record with no events.
What would break this evaluation: validators who pass all visible metrics can still fail through infrastructure problems that aren't detectable in advance. Geographic concentration of validator infrastructure creates correlated outage risk that doesn't show up in individual validator stats. Slashing events are also rarely predictable — they typically result from software bugs, key management errors, or edge cases in protocol upgrades. Past performance reduces probability but doesn't eliminate risk.
Timing Perspective
Now: Identify validators with public identities, on-chain performance history, and zero slashing events. Commission differences between 5–10% produce small yield variance. Focus on reliability, not rate minimization.
Next: If DVT adoption accelerates, the evaluation model changes. Worth tracking as a new dimension.
Later: Validator concentration is a known structural concern. If a small number of operators control a dominant share, effective validator choice narrows in ways that may require a different approach.
Boundary Statement
This covers the selection criteria — what variables matter and why. It doesn't address the tax treatment of staking rewards, the technical setup of running your own validator, or yield comparison between different staking protocols.
The right choice depends on your protocol, risk tolerance, and how actively you want to monitor ongoing performance. This is the static explanation. The operational tracking lives elsewhere.
