Delegating is not a neutral act. Two people staking the same amount into the same subnet can earn meaningfully different yields, because they picked different validators. Here is what actually drives that difference.
1. Take (commission)
A validator shares its emissions with delegators after deducting its take. A validator at 9% take passes through less than one at 5%, all else equal. Two nuances:
- Zero isn't automatically best. Running reliable validation infrastructure costs money; a sustainable take funds uptime, monitoring and upgrades. A race-to-zero validator that skimps on operations can cost you more in missed emissions than it saves in commission.
- Takes can change. A delegation is not a contract; check periodically that the terms you chose are still the terms you have.
2. Uptime and consistency
Emissions accrue every tempo, ~72 minutes. A validator that is offline, misconfigured, or slow to score miners earns less, and so do its delegators. You can read this in the data: a healthy validator shows steady emission capture over weeks, not sawtooth gaps. Consistency across time matters more than a spectacular week.
3. Size and stake position
Validator influence in a subnet scales with its stake. Larger validators enjoy stable positions and economies of scale; but concentration cuts the other way: a network where a handful of validators hold most stake is more fragile, and some subnets reward a broader validator set. There is also a subtler point: an enormous validator spread thin across every subnet may capture emissions less efficiently in a specific subnet than a focused operator with strong local positioning.
The compounding detail people miss
Small APY differences compound. At ~20 payouts per day, a 1-point difference in effective yield (take + uptime combined) is not a rounding error over a year. It is real TAO. Choosing a validator is a one-time decision with a recurring payoff; it deserves ten minutes of attention.
A practical checklist
- Take at or below the market norm, and stable over time.
- Weeks of uninterrupted emission capture in the subnets you care about.
- Meaningful but not dominant stake in those subnets.
- A team that communicates: infrastructure notes, incident reports, contact.
Or let the routing do it
This analysis has to be redone per subnet: the best validator for one subnet is rarely the best for all twenty in a diversified portfolio. That is why TaoScale strategies route each subnet's slice to a strong validator for that subnet automatically, and re-check as conditions change. You still hold your keys, you still see every delegation on-chain from your portfolio. The selection work is just done for you, without custody.