Station Labweek23 Reward Options

Why we need to keep payouts low:

• The expected stage of retrieval attestation / SPARK fraud detection won’t give enough confidence to pay out significant rewards

What we’re hoping to achieve:

• Disincentivize cheating, so that there would be no cheaters, which creates a nice experience for everyone

• Or: Make it so even if people cheat massively, the impact is low - it will only be unfulfilling for those who don’t cheat

Options for what to pay out:

• Low amounts of FIL (mainnet)
  • Pro:
    • The same coin that we will eventually pay out anyway
  • Con:
    • Even if we keep the reserve small (say 100 FIL) a cheater that can claim it all will still gets a noteworthy amount (99.9 FIL)
    • If we reward less, like 25FIL, with the current network size of 60 nodes this leads to very small .4FIL payouts per peer (assuming everyone is honest and performs the same amount of work)

• FIL (testnet)
  • Pro:
    • Similar enough to FIL mainnet
    • Doesn’t require architectural changes
    • We could scale payments by the amount of retrievals, instead of using a fixed size pot. This has a huge advantage: Cheating nodes won’t ruin the experience for honest nodes. They will just receive more, but honest nodes won’t receive less.
  • Con:
    • There is no value whatsoever associated, since everyone can get testnet FIL for free

• New token
  • Pro:
    • This token can eventually become valuable
  • Con:
    • This will require architectural changes
    • This deviates from the goal of paying out mainnet FIL eventually

Idea: If we pay testnet FIL and pay per retrieval and not a fixed amount per round, then the honest nodes’ experience won’t be ruined by cheaters