Proof of Stake

Every blockchain needs a way to agree on which transactions are valid without a trusted referee. The first answer was proof of work: make validators solve a computational puzzle, so cheating costs electricity. It worked — but burning energy at scale felt like a strange price for a digital ledger.

Proof of stake (PoS) replaced the puzzle with an economic bond. Instead of solving hashes, a validator stakes — locks up — a pile of coins as collateral. The protocol selects validators to propose and vote on blocks, weighted by how much they have at stake. A validator who tries to deceive the network gets slashed: part or all of their staked coins are destroyed. Cheat and you lose your deposit.

The key insight is that proof of work converts external energy into security, while proof of stake converts internal economic value. The cost of an attack is not the electricity bill but the market value of the coins destroyed in the process. Ethereum completed its switch to proof of stake in The Merge of September 2022 — one of the largest live migrations in software history, cutting the network's energy use by roughly 99.95 %.

The simulator below shows four validators, each with a different stake. The bar on the right of each row shows their selection probability — proportional to their stake share. Click Select validator to run one weighted draw and see who gets to propose the next block.

<div class="panel">
  <table id="vtable" aria-label="{{tbl_aria}}">
    <thead><tr><th>{{th_validator}}</th><th>{{th_stake}}</th><th>{{th_share}}</th><th></th></tr></thead>
    <tbody id="tbody"></tbody>
  </table>
  <div class="actions">
    <button id="select-btn" type="button">{{btn_select}}</button>
    <button id="add-btn" type="button" class="ghost">{{btn_add}}</button>
  </div>
  <div id="result" class="result" aria-live="polite"></div>
  <div class="log-label">{{log_label}}</div>
  <ul id="log" class="log"></ul>
</div>

* { box-sizing: border-box; }
body { font-family: system-ui, sans-serif; color: #222; margin: 0; }
.panel { max-width: 480px; }
table { width: 100%; border-collapse: collapse; font-size: .9rem; margin-bottom: .7rem; }
thead th { text-align: left; font-weight: 600; border-bottom: 2px solid #c5cdd6; padding: .3rem .4rem; font-size: .8rem; text-transform: uppercase; letter-spacing: .04em; color: #556; }
tbody td { padding: .38rem .4rem; border-bottom: 1px solid #e5e8ec; vertical-align: middle; }
.bar-wrap { background: #e5e8ec; border-radius: 4px; height: 10px; min-width: 80px; overflow: hidden; }
.bar { height: 10px; border-radius: 4px; transition: width .35s; }
.pct { font-size: .78rem; color: #556; min-width: 3.2em; display: inline-block; }
.slash-btn { font-size: .75rem; padding: .18rem .5rem; border: 1px solid #c92f3c; background: #fff; color: #c92f3c; border-radius: 5px; cursor: pointer; white-space: nowrap; }
.slash-btn:hover { background: #c92f3c; color: #fff; }
.slash-btn:disabled { opacity: .4; cursor: default; }
.actions { display: flex; gap: .5rem; flex-wrap: wrap; margin-bottom: .6rem; }
button { font: 600 13px system-ui, sans-serif; padding: .42rem .9rem; border: 1px solid #1d3557; background: #1d3557; color: #fff; border-radius: 7px; cursor: pointer; }
button.ghost { background: #fff; color: #1d3557; }
.result { font-size: .95rem; font-weight: 600; min-height: 1.5em; padding: .35rem .6rem; border-radius: 6px; transition: background .2s; margin-bottom: .5rem; }
.result.selected { background: #d4f5e0; color: #0a5c2a; }
.log-label { font-size: .75rem; font-weight: 600; text-transform: uppercase; letter-spacing: .05em; color: #778; margin-bottom: .3rem; }
.log { list-style: none; padding: 0; margin: 0; font-size: .82rem; max-height: 90px; overflow-y: auto; color: #445; }
.log li { padding: .12rem 0; border-bottom: 1px solid #f0f2f4; }
.log li:last-child { border-bottom: none; }
tr.highlight td { background: #fff8c5; transition: background .6s; }

// Code not found

Try Slash on the leading validator: their stake drops, redistributing probability to the rest. Add validator introduces a new participant. Notice how adding small stake barely moves the needle for the giants — this is why concentration of stake is a real concern, and why many protocols impose caps or diminishing returns.

Replacing energy with economics sounds clean, but it moves the hard problems rather than eliminating them.

• Nothing-at-stake: in early PoS designs, a validator caught a fork could safely vote on both branches at no cost. Proof of work miners physically cannot mine two chains at once; PoS validators could. The solution is slashing — voting on two conflicting blocks with the same key gets your deposit destroyed.
• Long-range attacks: an attacker who once held a large stake could reconstruct an alternative history from far in the past, since old keys cost nothing to sign with. Defenses include weak subjectivity checkpoints — trusted snapshots beyond which the chain cannot be reorged.
• Stake concentration: if large holders can always win more stake, the rich get richer and the network centralizes. Many protocols add delegation, random sampling among a committee, or diminishing return formulas.
• Validator liveness vs. safety: Byzantine fault-tolerant (BFT) consensus requires ⅔ of stake to be honest for safety and ⅔ to be online for liveness. Ethereum's Casper protocol separates finality (BFT vote over checkpoints) from block proposal (a weaker lottery), accepting that the two properties can occasionally conflict.

The computational hardness of PoS is mostly absent — selection is a weighted random draw, trivially polynomial. The difficulty is entirely in the game-theoretic design: making rational defection unprofitable rather than computationally infeasible. This is a fundamentally different kind of security argument than the one behind proof of work.

Proof of stake is now the dominant consensus mechanism for new blockchains, and its influence extends well beyond cryptocurrency:

• Ethereum 2.0: after The Merge, Ethereum runs on Casper PoS with ~900 000 active validators staking 32 ETH each. Finality is guaranteed in about 12 minutes under normal conditions.
• Delegated PoS (DPoS): networks like EOS and TRON let token holders vote for a fixed set of block producers. Governance becomes explicit but concentration risk is higher.
• Nominated PoS (NPoS): Polkadot uses an election algorithm (Phragmén's method) to assign nominators to validators in a way that maximizes the minimum stake behind any validator — a fairness criterion.
• Sharding: Ethereum's roadmap uses PoS committees to attest to shard blocks in parallel, with slashing ensuring committees can't collude cheaply. The randomness needed for committee assignment is itself a hard distributed-systems problem.
• Mechanism design: the slashing-plus-reward structure is a classic incentive-compatibility result — telling the truth (following the protocol) is a dominant strategy when lying costs more than it gains.

Proof of stake is a reminder that "hard to cheat" does not have to mean "hard to compute." Instead of making an attack computationally infeasible, PoS makes it economically suicidal — attack the chain and you destroy the very asset that gave you influence.

The shift from energy to economics does not remove the difficulty; it relocates it. The math of weighted selection is trivial. The depth is in the protocol design: preventing long-range attacks, thwarting nothing-at-stake incentives, keeping validator sets decentralized, and guaranteeing finality under a Byzantine minority. Those are problems for game theory and distributed systems, not for polynomial-time algorithms.

Every time you read about a network "going green" by switching to PoS, you are watching that trade — computational hardness swapped for incentive design. Whether the bargain holds depends not on cryptography alone but on whether the economic assumptions survive contact with the real world.