How SolPilled measures Solana

Every number on this site is read, computed or stated. This page shows which is which, every formula behind the homepage, where the data comes from and what stops a bad reading from going live.

Updated Data as of 2026-11-16 00:42 UTC 7 sources
Short answer

Every number on SolPilled is read (from the Solana network or a public API, with the time), computed from reads by a formula on this page, or stated by a document, with its source. Yearly figures are per calendar year, compounded per epoch. The latest reading, 2026-11-16 00:42 UTC, puts supply growth at 3.37% a calendar year and staking at a typical validator at 4.79%, so a staker's share of all SOL grows +1.37% a year.

Three kinds of number

Read

Taken from the Solana network (the public JSON-RPC) or a public API at a stated time. Example: the network's inflation rate, 3.556% at 2026-11-16 00:42 UTC. Pages mark numbers read from the network that day in yellow.

Computed

Arithmetic on reads, by a formula written on the page or below. Example: the change in a staker's share of all SOL, +1.37% a year. We never call arithmetic a measurement.

Stated

Quoted from an app, a protocol document or a company, with its source and the date we read it. Example: the 5,000-lamport fee per signature. We never repeat an app's advertised yield as if we measured it.

One SOL is 1,000,000,000 lamports. Small amounts are given in SOL with the lamports in brackets, for example the token-account deposit: 0.00148844 SOL (1,488,440 lamports). Times are UTC.

Every homepage number and its formula

These are the numbers at the top of the homepage, with the value from the latest reading (2026-11-16 00:42 UTC). The formulas are the ones the site's collector runs, and the readings are kept day by day in the daily readings dataset.

NumberKindHow it is madeLatest
Network inflation ratereadgetInflationRate (total). Per protocol year of 200 ms slots, not per calendar year3.556%
Epoch length and epochs a yearcomputedmean length of the last completed epochs, from block times at the epoch boundaries (getBlockTime); if those reads fail, 432,000 slots × measured slot time; N = 8,760 hours ÷ measured epoch length25.76 h; 340.1
New SOL per epochcomputedi = network inflation rate × 432,000 ÷ 157,784,629.968 (new SOL per epoch, as a share of total supply)0.00974% of supply
Staker reward per epochcomputedg = i × total supply ÷ active stake (new SOL per epoch per staked SOL, before commission; MEV tips and shared block rewards are extra)0.01447%
Supply growth in a calendar yearcomputed(1 + i)^N − 13.37%
Staking yield before commissioncomputed(1 + g)^N − 15.04% a year
Staking yield at a typical validatorcomputed(1 + g × (1 − commission))^N − 1, shown before commission and at the most common commission among validators charging under 100%4.79% a year (at 5%)
Change in a staker's share of all SOLcomputed(1 + staking yield) ÷ (1 + calendar-year inflation) − 1, the change in a staker's share of all SOL over a year+1.37% a year
Change in an unstaked holder's sharecomputed1 ÷ (1 + calendar-year inflation) − 1, the change in the share of all SOL held by SOL that is not staked−3.26% a year
Share of SOL stakedcomputedactive stake (getVoteAccounts, current validators) ÷ total supply (getSupply)67.3%
Stake-weighted average commissioncomputedΣ(stake × commission) ÷ Σ stake over current validators. Published, but not what a delegator pays25.94%
Validators that together pass one third of stakecomputedfewest validators, largest first, whose stake together exceeds one third of all stake, delinquent stake included18
Validators voting and delinquentreadgetVoteAccounts: current and delinquent lists662 and 38
Block (slot) timecomputedseconds ÷ slots over the last 30 one-minute performance samples215 ms (target 200 ms)
User transactions a secondcomputednon-vote transactions ÷ seconds over the same samples (since Alpenglow, votes are not transactions)1,771
Hours left in the epochcomputedslots left in the epoch × measured slot time4.9 h
Token-account depositreadgetMinimumBalanceForRentExemption(165), the size of a classic token account1,488,440 lamports
Empty-account depositreadgetMinimumBalanceForRentExemption(0)650,240 lamports
Fee per signaturestatedProtocol constant, from the Solana fee documentation; half is burned, half goes to the block producer5,000 lamports
SOL to send USDC to a new walletcomputedtoken-account deposit + one 5,000-lamport signature fee (the self-funded case; some wallets offer gasless sends paid from the token)0.00149344 SOL
Jito tip, median landedreadJito tip floor API, landed_tips_50th_percentile1,444 lamports
Dollar amounts of costscomputedlamports ÷ 1,000,000,000 × the SOL/USD rate from the DefiLlama coins API, read with the rest. Used for conversion only$71.71 per SOL

The typical commission is the most common commission among validators that charge less than 100%: 5% today (265 validators), then 0% (220). We do not use the stake-weighted average (25.94%) because 23% of all stake sits with 61 validators that charge 100%: operators staking their own SOL, which nobody delegates to for a reward.

Real yield, worked through

A staking reward is paid in new SOL, and new SOL dilutes everyone. So the question is how your share of all SOL changes, not how many SOL you gain. One more trap first: the network's inflation rate is per protocol year, a year of exactly 200 ms slots (157,784,629.968 slots). Slots take about 215 ms, so a calendar year holds about 340 epochs, not 365, and every calendar-year figure is lower than the raw rate suggests. SolPilled works per epoch and compounds over the measured number of epochs in a calendar year. With the reading of November 16, 2026:

  1. Network inflation rate (read): 3.556% per protocol year. The schedule started at 8%, falls 15% a year and stops at 1.5%.
  2. New SOL per epoch (computed): 3.556% × 432,000 ÷ 157,784,629.968 = 0.00974% of supply.
  3. Staker reward per epoch (computed): that × 637.62M SOL total supply ÷ 428.97M SOL active stake = 0.01447%. New SOL goes only to staked SOL, which is why the reward is higher than inflation.
  4. Epochs in a calendar year (computed): 8,760 hours ÷ 25.76 hours = 340.1.
  5. Supply growth in a calendar year (computed): (1 + 0.00974%)340.1 − 1 = 3.37%.
  6. Staking yield at a 5% commission (computed): (1 + 0.01447% × 0.95)340.1 − 1 = 4.79% (before commission: 5.04%).
  7. Change in a staker's share (computed): (1 + 4.79%) ÷ (1 + 3.37%) − 1 = +1.37% a year.
  8. Change in an unstaked holder's share (computed): 1 ÷ (1 + 3.37%) − 1 = −3.26% a year.

We checked the per-epoch figure against what validators were actually paid: for three validators that charge 100% (so the whole reward stays with the stake), the reward divided by their stake in one epoch was 0.0141% to 0.0148% (getInflationReward, 2026-11-10).

This counts the inflation reward only. MEV tips that validators share add a little (Jito's tip floor and the staking pages measure them); priority fees go to the block producer and are not shared with delegators by default. Liquid staking tokens are measured differently: from the pool's own exchange rate across epochs, annualized with the measured epoch length, never with an advertised APY. Those numbers are on the staking pages.

Where the data comes from

  • The Solana network, through the public JSON-RPC (api.mainnet-beta.solana.com, solana-rpc.publicnode.com, with a second public endpoint as fallback): getEpochInfo, getSupply, getInflationRate, getVoteAccounts, getRecentPerformanceSamples, getMinimumBalanceForRentExemption. One request at a time, a few dozen calls per run, no keys.
  • Jito's public API for the tip floor (what landed tips paid).
  • DefiLlama's open API for app deposits, fees, DEX volume and stablecoins on Solana, and for the SOL/USD rate used to convert costs. DefiLlama is a soft source: if it fails, the previous value stays.
  • Documents for stated facts: the Solana documentation (read from its public source repository when solana.com turns automated readers away), SIMDs (Solana improvement documents), feature-activation accounts on chain, and apps' own help pages and terms, each linked where it is used.

We do not use sources we cannot check. The public RPC's getRecentPrioritizationFees returns zeros, so fee levels are measured by sampling recent blocks instead. One staking aggregator's public API returned stale exchange rates when we tested it, so pool rates are read from the pools' own accounts. The datasets on the data page (14 today) each list their own sources, method and caveats.

Time, epochs and the 2026 changes

Several numbers that most pages still repeat changed in 2026, so we read them instead of typing them:

  • Slot time: the target fell from 400 ms to 200 ms in four steps between 2026-08-19 and 2026-10-08. Measured: 215 ms over the last 30 minutes.
  • Epoch length: an epoch is 432,000 slots, and recent epochs lasted about 25.76 hours each (measured from block times at the boundaries), about 340 epochs a calendar year, not two days. Anything annualized per epoch uses the measured length, never a fixed "epochs per year". Native unstaking completes at the next epoch boundary: 4.9 hours away at the latest reading.
  • Votes: since Alpenglow went live on 2026-10-21, validator votes are no longer transactions. "Transactions a second" here counts user transactions only; older figures that included votes are not comparable.
  • Account deposits: the token-account deposit is 1,488,440 lamports (0.00148844 SOL), down from 2,039,280 (the old figure many pages still quote). The RPC has no historical parameter, so we never apply today's figure to an old date.

The gates

A reading only replaces the last one when it passes its gate. For the homepage numbers the gate checks that total supply is between 400 million and 1 billion SOL, inflation between 1% and 9%, active stake above 200 million SOL and below total supply, between 300 and 5,000 validators, the token-account deposit between 100,000 and 5,000,000 lamports, slot time between 100 and 1,000 ms, and the staking reward between 1% and 15%. Each dataset has its own gate (ranges, row counts and a comparison with the previous file).

  • A failing gate keeps the last good file. The page then shows its older reading, with its time, and dataset pages say "older snapshot" when a reading is more than 8 hours old.
  • Soft sources (anything other than the Solana RPC) fall back to their previous value instead of holding the whole reading.
  • The build itself refuses to publish a dataset that breaks the data contract, a page title over 60 characters, a dead internal link, or a page whose machine-readable server answers differ from the page.

What we do not publish

SolPilled does not discuss, chart or predict token prices and makes no picks. The SOL/USD rate it reads is used only to show a cost in dollars. We do not rank staking tokens or validators with words like "best" or "safest": we publish measured columns, dated, and let them be sorted. We do not publish a validator leaderboard or a live dashboard; <a href="/network/">the network page</a> is a dated fact sheet with a method. SolPilled never connects to a wallet, never asks for a seed phrase and builds nothing that signs a transaction. Its calculators compute from numbers already on the page.

Corrections

If a number, date or claim here is wrong, tell us. We check every report against a primary source, fix verified errors, change the page's Updated date and note any correction that changes an answer. We do not delete the record of a mistake. Your report is stored with the time and your country; your email is optional and used only to reply by hand. This site sends no automatic email.

Questions people ask

How does SolPilled calculate real staking yield?

It divides one plus the staking yield by one plus the supply growth over the same calendar year and subtracts one: the change in your share of all SOL. On November 16, 2026 that was (1 + 4.79%) ÷ (1 + 3.37%) − 1 = +1.37% a year at a 5% commission validator.

Why is Solana's inflation rate higher than the supply growth you show?

The network reports 3.556% per protocol year, a year of exactly 200 ms slots. Slots take about 215 ms, so a calendar year holds about 340 epochs, not 365, and supply grows 3.37% in a calendar year. SolPilled shows calendar-year figures.

How often are the numbers updated?

Every two hours: the site reads the network, checks the reading against its gates and rebuilds. The current reading is from 2026-11-16 00:42 UTC.

How do I report a wrong number?

Use the corrections form on this page. Every report is checked against a primary source; a verified error is fixed, the page's Updated date changes, and a correction that changes an answer is noted on the page.

Sources

  1. Solana JSON-RPC methods (solana.com docs)
  2. Fees on Solana (solana.com docs)
  3. Reduced slot times (solana.com)
  4. Alpenglow (solana.com)
  5. Reduced rent (solana.com)
  6. Jito tip floor API
  7. DefiLlama API (api.llama.fi)