Understand › Staking

Staking your FLR: what a validator really pays

A Flare validator’s yield is made up of two rewards, each reduced by its own fee. Here is how to read them, and why the highest figure is not always the best choice.

When you stake FLR on the P-chain with a validator, you do not receive one reward, but two. They come from two different pools, are calculated differently and each bears its own fee. That is what makes yields so hard to compare from one site to another, and that is what this article untangles.

Over the epochs 443 to 450, the median net yield of Flare validators is 9.11% per year, for 98 validators with more than one million FLR. This figure, like all those underlined with dots in this article, is recalculated every epoch from the Flare Foundation’s public files.

Two rewards in a single yield

Flare distributes its rewards in epochs of 3.5 days, i.e. 104.29 epochs per year. Each epoch, a staker receives:

  1. The “staking” share. This is a dedicated pool, split among validator nodes by the Flare Foundation’s public scripts. It pays exactly the same per FLR with every eligible validator, apart from the node’s fee: currently 2.83% per year at the median.
  2. The “mirror” share. A Flare validator is also, almost always, a data provider for the protocol: it publishes prices (the FTSO) and attests data from other blockchains (the FDC). For this work, it receives rewards, and part of them is passed on to those who staked with it, pro rata. This share is what makes all the difference between two validators.

The yield shown by Flarescope is the sum of the two, as a simple rate, with no compounding.

A staker's yield adds up two shares. The staking share comes from the staking pool, minus the P-chain node fee. The mirror share comes from the data provider's FTSO and FDC rewards, minus its FTSO fee. Staking pool same rate everywhere FTSO, FDC rewards of the data provider − node fee P-chain (often 20%) − FTSO fee of the provider Staking share Mirror share Net yield shown
Two rewards, each with its own fee. The mirror share is what makes the difference between two validators.

Two fees, to be read separately

Each of the two shares has its own fee, and the same validator can set different rates:

  • the P-chain node fee applies to the staking share. Most validators set it at 20%, some at 10%;
  • the data provider’s FTSO fee applies to the mirror share. It is currently 20% for every validator tracked.

Both are already deducted from the yields shown on this site. A low fee is therefore not necessarily an advantage: a validator at 10% that publishes poor data can pay less than a flawless validator at 20%.

Key point

Compare net yields, not fees. A difference of one point of yield is worth 10,000 FLR per year for one million FLR staked.

Why small validators pay more

This is the most counter-intuitive result: the bigger a validator, the less it pays per FLR. The reason lies in how the protocol measures a data provider’s weight. According to the distribution files, this weight follows the formula:

weight = (5 × P-chain stake + delegated WFLR)0.75

Mirror rewards are roughly proportional to this weight. But the 0.75 exponent makes weight grow more slowly than stake: a validator ten times bigger receives only about 5.6 times more rewards, and shares them among ten times more FLR. Each FLR then receives about 44% less on the mirror share.

A validator ten times bigger receives about 5.6 times more mirror rewards, which it shares among ten times more FLR: each FLR receives about 44% less. Small 10 times bigger Stake × 10 Mirror rewards × 5.6 Per FLR staked − 44% provider weight ≈ (5 × stake + WFLR)^0.75
Example with no delegated WFLR: the weight, and so the rewards, grow more slowly than the stake. The “per FLR” bars use a different scale from the others.
0.0%2.5%5.0%7.5%10.0%12.5%10 M100 M1 B0x04cfe617: 154.5 M FLR, 9.61%0x7e74d30e: 1002.4 M FLR, 6.93%0x802c64fc: 884.2 M FLR, 7.60%0x882ee9ce: 27.5 M FLR, 12.26%0x9e5ef16d: 31.3 M FLR, 9.92%0xc4019e18: 28.5 M FLR, 11.54%0xdd8fe872: 14.4 M FLR, 11.45%1FTSO: 52.2 M FLR, 11.14%4DadsFTSO: 149.6 M FLR, 8.74%A-FTSO: 199.4 M FLR, 9.03%AFOracle: 827.7 M FLR, 7.20%African Proofs: 115.6 M FLR, 9.06%Aimlezz: 120.3 M FLR, 10.05%AlphaOracle: 628.6 M FLR, 7.74%Ankr: 295.1 M FLR, 1.08%Aternety: 311.3 M FLR, 8.86%Atlas TSO: 612.7 M FLR, 7.48%AU: 886.2 M FLR, 7.56%Aureus Ox: 575.7 M FLR, 7.81%Bifrost Wallet: 852.7 M FLR, 6.73%Burst FTSO: 54.2 M FLR, 10.94%Bushido FTSO: 322.0 M FLR, 8.98%CashLab: 108.1 M FLR, 10.36%Catenalytica: 438.8 M FLR, 8.48%Chainbase Staking: 335.0 M FLR, 8.48%Comfy Nodes: 59.5 M FLR, 6.74%Cottage Nodes: 50.5 M FLR, 9.09%D20 Oracle: 84.0 M FLR, 11.58%DataVector: 188.9 M FLR, 8.77%Defi Oracles: 31.3 M FLR, 8.89%Digital Dynamix: 232.2 M FLR, 8.38%Encode Club: 197.5 M FLR, 8.69%Envision: 191.5 M FLR, 9.56%EvolveFTSO: 179.3 M FLR, 9.13%Flare Bank: 153.8 M FLR, 9.42%Flare Beacon: 188.5 M FLR, 8.88%Flare Dienst: 66.0 M FLR, 9.17%Flare En Espanol: 31.0 M FLR, 12.01%Flare Oracle: 213.2 M FLR, 8.84%Flare Sensei: 465.6 M FLR, 8.22%Flare.Space: 1060.4 M FLR, 6.95%FlareBase: 155.7 M FLR, 9.75%FlareBus: 501.6 M FLR, 7.98%FlareFi: 281.1 M FLR, 8.43%FlareForward: 70.8 M FLR, 10.57%FlareWatch: 228.9 M FLR, 9.27%Flaris: 179.1 M FLR, 9.84%FocusTSO: 13.6 M FLR, 3.78%FTSO Asia: 44.8 M FLR, 10.61%FTSO EU: 127.2 M FLR, 9.15%FTSO London: 75.0 M FLR, 10.18%FTSO Paris: 319.2 M FLR, 9.03%FTSO Plus: 137.8 M FLR, 9.74%FTSO UK: 10.3 M FLR, 2.75%FTSO Yggdrassil: 88.7 M FLR, 10.41%FTSOBest: 77.0 M FLR, 10.36%FTSOCAN: 102.5 M FLR, 9.70%G1Flare818: 39.1 M FLR, 12.26%HEWG: 136.8 M FLR, 9.86%Honest Systems: 111.2 M FLR, 10.24%HP/Mana Nodes: 158.5 M FLR, 9.63%InfStones: 413.0 M FLR, 8.14%Juice Nodes: 151.3 M FLR, 9.40%Kiln: 645.6 M FLR, 7.22%Knot Nodes: 100.9 M FLR, 9.55%Last Oracle: 715.9 M FLR, 6.87%LightFTSO: 73.7 M FLR, 10.44%Linden Services: 172.1 M FLR, 8.94%Luganodes: 296.1 M FLR, 8.64%Mickey B Fresh: 67.2 M FLR, 9.64%Mirhollio Core: 106.1 M FLR, 10.58%Nansen: 89.6 M FLR, 8.84%Next Finance Tech: 529.4 M FLR, 5.24%NORTSO: 438.2 M FLR, 8.06%Oracle Daemon: 233.4 M FLR, 9.03%Pingu: 62.3 M FLR, 10.90%PRICEKRAKEN: 496.3 M FLR, 8.25%Quicknode: 258.9 M FLR, 8.45%Radroot: 9.3 M FLR, 3.66%Resonance: 256.6 M FLR, 9.14%Scintilla: 135.5 M FLR, 6.29%SenseiNode: 74.5 M FLR, 10.11%Seraphim FTSO: 119.4 M FLR, 10.49%Solarius: 169.0 M FLR, 9.27%SparkDEX: 29.7 M FLR, 9.81%Sparky: 58.2 M FLR, 10.67%Stakecore: 73.6 M FLR, 10.38%sToadz FTSO: 171.3 M FLR, 9.31%Sun-Dara: 207.5 M FLR, 8.68%Tailwind FTSO: 39.2 M FLR, 10.10%TempestFTSO: 298.0 M FLR, 7.51%True FTSO: 39.6 M FLR, 9.24%uGaenn: 177.3 M FLR, 9.60%Ugly Kitty: 345.0 M FLR, 8.59%Use Your Spark: 129.5 M FLR, 9.70%WitterFTSO: 66.2 M FLR, 9.34%Wonderftso: 39.2 M FLR, 7.89%Xdrops Oracle: 186.0 M FLR, 6.23%
Net yield (% per year) by validator stake, logarithmic scale, epochs 443 to 450. One dot per validator. Source: Flarescope calculation from the Flare Foundation’s public files.

The scatter plot shows it: the trend clearly slopes down to the right. The spread around the trend comes from the quality of the data provider’s work and from its fees.

What your own stake changes

The same mechanism applies to you. By staking with a validator, you increase its stake, so you slightly dilute its mirror share, for it as for you. For a small amount, the effect is negligible; for several million FLR with a small validator, it becomes noticeable. The simulator calculates this dilution before showing an estimate.

The staking share, on the other hand, is not diluted by your stake, with one exception: a single entity cannot weigh more than 5% of the network’s total stake in this calculation. Beyond that, its staking share is reduced. No validator reaches this threshold today.

The rules that limit your choice

Three rules, taken from the official documentation, govern staking:

  • 50,000 FLR minimum per stake;
  • 2 weeks minimum duration;
  • 15 times the self-bond: a node cannot receive more than 15 times what its validator has put in itself. A validator in high demand can therefore be full.
Example: a validator that has put in 10 million FLR of its own can receive a total stake of 15 times that amount, i.e. 150 million. If it already has 120 million, 30 million of capacity remain. Whatever the calculation, a node never exceeds 300 million FLR. Example: self-bond of 10 M FLR ↑ self-bond 10 M stakes received 110 M capacity 30 M cap 15 × 10 M = 150 M Three limits on a stake • at least 50,000 FLR and 2 weeks • no more than 15 times the validator's self-bond per node • 300 M FLR at most per node (FIP.16)
A validator in high demand can be full: the comparison shows each one’s remaining capacity.

The comparison shows, for each validator, the remaining capacity and the end date of its nodes, based on the P-chain state read every 30 minutes.

What about eligibility?

A data provider that does not meet the protocol’s minimum conditions during an epoch (notably its node’s availability and its participation in the data protocols) can lose its rewards for that epoch. So can its stakers. Flarescope counts these epochs as zero in the average, which penalises irregular validators: this is deliberate, because it is what you would have received.

In short

  • A validator’s yield is a staking share that is almost the same everywhere, plus a mirror share that varies from one validator to another.
  • Small validators mechanically pay more per FLR, but they have less capacity.
  • The best past yield is no guarantee: also look at consistency, remaining capacity and the end date of the nodes.

The details of the calculation, its sources and its limits are described on the Method page.