heimdall-stake: stake function and splitting-gain report (Rust, zero deps) #9

Merged

libby merged 2 commits from libby/heimdall-stake into main 2026-11-06 04:34:33 +00:00

Conversation 1 Commits 2 Files changed 9 +572

libby commented 2026-11-05 11:43:15 +00:00

Member

Copy link

New package packages/heimdall-stake, beside heimdall-delegation. weight = 1 + log_b(amount/unit) per Meta, in integer cents and integer micro-weights (1,000,000 = 1.0) with an integer-only logarithm (deterministic across platforms), plus Heimdall's floor, per-identity cap and minimum amount per identity, and a splitting-gain report (splitting_gain, splitting_curve, most_profitable_split). README: first call, formula with Meta's example (100 EUR is 3, 10,000 EUR is 5, both tested exactly), prior art (quadratic voting, Lalley-Weyl, Sybil; cited from background, marked as not re-fetched), and the note that the estate rule normalizes above and the sybil rate is a harness adversary. All parameters are placeholders until the estates rule.

Tests: 11 property/behaviour tests pass (cargo test in the package; 0 unit tests, 0 doctests): Meta's example, exact powers for other bases and units, integer log vs f64 within 2 micro-weights over 20,000 draws, monotone, concave above the unit, zero at and below the floor, cap never exceeded and binds, deterministic and total over i64::MIN..MAX, bad parameters are errors, the split report equals a hand sum, and the minimum makes splitting stop paying.

Choices to confirm: floor is exclusive (an amount at or below it weighs 0; the brief said both "below" and "at/below"); per_identity_cap caps weight, not amount; floor and minimum both zero a weight and may want to be one parameter; base is an integer of at least 2.

A finding worth Heimdall's eyes: the log formula's +1 rewards splitting (100 EUR as ten identities of 10 EUR weighs 20 against 3), and the minimum prices it as a cliff, not a slope. With a 50 EUR minimum, 100 EUR in two identities still gains (5.40 vs 3.00); in three it weighs 0. The minimum has to sit near a typical budget to deter splitting.

Not merged.

🤖 Generated with Claude Code

New package `packages/heimdall-stake`, beside `heimdall-delegation`. `weight = 1 + log_b(amount/unit)` per Meta, in integer cents and integer micro-weights (1,000,000 = 1.0) with an integer-only logarithm (deterministic across platforms), plus Heimdall's floor, per-identity cap and minimum amount per identity, and a splitting-gain report (`splitting_gain`, `splitting_curve`, `most_profitable_split`). README: first call, formula with Meta's example (100 EUR is 3, 10,000 EUR is 5, both tested exactly), prior art (quadratic voting, Lalley-Weyl, Sybil; cited from background, marked as not re-fetched), and the note that the estate rule normalizes above and the sybil rate is a harness adversary. All parameters are placeholders until the estates rule. Tests: 11 property/behaviour tests pass (`cargo test` in the package; 0 unit tests, 0 doctests): Meta's example, exact powers for other bases and units, integer log vs f64 within 2 micro-weights over 20,000 draws, monotone, concave above the unit, zero at and below the floor, cap never exceeded and binds, deterministic and total over i64::MIN..MAX, bad parameters are errors, the split report equals a hand sum, and the minimum makes splitting stop paying. Choices to confirm: floor is exclusive (an amount at or below it weighs 0; the brief said both "below" and "at/below"); `per_identity_cap` caps weight, not amount; floor and minimum both zero a weight and may want to be one parameter; base is an integer of at least 2. A finding worth Heimdall's eyes: the log formula's +1 rewards splitting (100 EUR as ten identities of 10 EUR weighs 20 against 3), and the minimum prices it as a cliff, not a slope. With a 50 EUR minimum, 100 EUR in two identities still gains (5.40 vs 3.00); in three it weighs 0. The minimum has to sit near a typical budget to deter splitting. Not merged. 🤖 Generated with [Claude Code](https://claude.com/claude-code)

libby added 1 commit 2026-11-05 11:43:15 +00:00

heimdall-stake: stake function and splitting-gain report (Rust, zero deps) ... 9adee90499

weight = 1 + log_b(amount/unit) in integer cents and micro-weights with an integer-only
log, plus floor, per-identity cap and minimum amount per identity; splitting_gain/curve
report what splitting a budget over k identities does. 11 property tests.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

libby commented 2026-11-05 12:43:12 +00:00

Author

Member

Copy link

Maintainer sanity run at head 9adee904: cargo test --offline in packages/heimdall-stake, 11 pass, 0 fail (0 unit, 0 doc). Forge CI: 6 pending, 3 skipped (workflows waiting for a runner; none has ever started on this repo). Not merged, Heimdall to review.

Maintainer sanity run at head 9adee904: `cargo test --offline` in packages/heimdall-stake, 11 pass, 0 fail (0 unit, 0 doc). Forge CI: 6 pending, 3 skipped (workflows waiting for a runner; none has ever started on this repo). Not merged, Heimdall to review.

libby added 1 commit 2026-11-06 01:31:13 +00:00

heimdall-stake: sybil note and golden vectors ... 6f68341edd

README: the constant 1 in weight = 1 + log_b(amount/unit) is paid once per identity, so splitting
always pays until the log goes negative and no base fixes it; the defence lives outside the function.
Golden vectors (single and split k=1,2,10,100, three bases) computed with 60-digit decimal ln,
pinned by tests/golden.rs.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>

libby force-pushed libby/heimdall-stake from 6f68341edd to 84f932ec50 2026-11-06 02:31:09 +00:00 Compare

libby merged commit b88beddecb into main 2026-11-06 04:34:33 +00:00

libby referenced this pull request from a commit 2026-11-06 04:34:35 +00:00

Merge pull request 'heimdall-stake: stake function and splitting-gain report (Rust, zero deps)' (#9) from libby/heimdall-stake into main

Sign in to join this conversation.

Reviewers

No reviewers

Labels

Clear labels

No items

No labels

Milestone

Clear milestone

No items

No milestone

Projects

Clear projects

No items

No project

Assignees

Clear assignees

No assignees

1 participant

Notifications

Subscribe

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference

GameCult/CultLib!9

No description provided.