Five rounds and an anchor: how the price finds itself
D17 blog · 02 · price discovery
Most token sales pick a price and hope. The creator guesses what the market will bear; the market punishes the guess in one direction or the other. Price too low and insiders scoop the discount; too high and the sale dies quietly.
A D17 launch does not hard-code one final sale price. It discovers the sale outcome, in public, over five timed rounds. The creator still chooses the token allocation, minimum anchor requirements, and schedule before deployment.
Round one is the vote
Round 1 is the anchor round. Committers put WETH in against a fixed slice of the sale allocation, and when the round and its refund window close, the ratio of money committed to tokens allocated becomes the launch's anchor price. The creator chose the allocation and minimums; the resulting ratio comes from actual commitments.
Pulling out of round 1 costs a fixed 1%, paid into the launch's pool. That small number matters: it means nobody can pump the anchor with a huge commitment and withdraw it for free at the last second.
Every later round is held to the anchor
Each subsequent round offers its own allocation of tokens, with a target: the WETH needed to sell that allocation at the anchor price. A round that beats its target sells everything, at a higher price. A round that falls short sells only what its money covers at the anchor price - never more.
That rule now applies to all four later rounds. No round can sell tokens below the price round 1 established. The sale's economics are set by its earliest, most-committed participants - not by the creator's optimism and not by whoever shows up last.
Because the schedule is fixed and published before round 1 opens - allocations per round, lengths, refund windows - there's no lever for anyone to pull once price discovery begins. The creator watches like everyone else.
Round five takes what's left, at a fair price
Real sales are lumpy. Some rounds oversubscribe, some fall short. Allocation that rounds 2–4 didn't place rolls into round 5, which gets one last chance to sell it - at or above the anchor price.
What round 5 can't sell at that price isn't handed out at a discount, and it isn't given to anyone. It's burned, along with the matching share of the liquidity allocation, so the pool's opening price follows what buyers actually paid. (In the earlier version of the protocol, round 5 handed its whole pool to whoever committed, at whatever price - a gift to a last-minute sniper. That door is closed.)
Why this beats a fixed price
- No hidden presale price inside the round mechanism. Any creator allocation is a separate, capped, visible part of the supply split - and it vests over six months.
- No stale guess. A price set weeks before launch can't respond to the market. An anchor set by round 1 is, by construction, current.
- No bargain bin at the end. The last round is held to the same floor as the others.
- The anchor requirements are explicit. If round 1 does not meet the published minimum raise and minimum anchor price, the launch fails and remaining commitments can be reclaimed in full, without penalty. Participants still pay their own transaction gas.
The quiet radical idea here is that a launch's price is a finding, not a setting. Five rounds is how long D17 gives the market to speak - and the contract writes down exactly what it heard.
The mechanics: 5 fixed rounds; round 1 sets the anchor price; rounds 2–5 sell min(allocation, raised ÷ anchor price) tokens; unsold allocation from rounds 2–4 rolls into round 5; whatever round 5 cannot sell at the anchor is burned at finalization, and the LP allocation is scaled to the share sold. A launch that misses its round-1 anchor requirements provides a permanent penalty-free refund path for remaining commitments.