fix #266: retail run-rate 800 branch is exact-equality sentinel, not a cap

Raw byte decode of MovementSystem::GetRunRate (0x006b0950, PDB-paired
binary): fild skill; fcom [800f]; fnstsw; test ah, 0x44; jp general —
the C2/C3 parity idiom whose 18/4 fall-through executes ONLY at
skill == 800 exactly. ACE read this as >= 800 ('max run speed?') and
Campaign P P1 inherited that misread when BN dropped the arithmetic,
flat-lining every maxed character at 4.5 (retail-true ~3.70, +21%) and
erasing the vitae differential (both 10200 and 15225 sat above 800).

The [stat-chain] live capture proved the enchant chain correct end to
end (vitae 0.67 -> eff run 10200 -> controller), isolating the formula.
General path byte-verified: (loadMod*(skill/(skill+200)*11)+4)/scaling/4.
InqMaxRunRate's skill=9999 probe gets ~3.6961, not 4.5.

Golden tests pin the 799/800/801 straddle and the maxed-skill vitae
differential; pseudocode doc §6 carries the decode plus a do-not-
reimport-ACE warning. Complete Release suite: 10,025 passed / 5 skips.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This commit is contained in:
Erik 2026-07-30 19:22:41 +02:00
parent bd3ade625f
commit 61e959169b
5 changed files with 85 additions and 28 deletions

View file

@ -156,22 +156,30 @@ first, don't guess the format string.
## #266 — Local player faster than a comparable retail character
**Status:** OPEN — needs a controlled comparison capture
**Status:** CLOSED 2026-07-30 — root cause: ACE-inherited `>= 800` misread
of retail's exact-equality run-rate sentinel
**Severity:** HIGH (matrix live gate 2026-07-30; core speed parity)
**Component:** movement / stat chain (Campaign P P1) / outbound speed
**Component:** movement / `MovementSystem.GetRunRate`
**Symptom (user report):** side-by-side, the acdream character runs
visibly faster than a character on retail. Confounds to rule out first:
(a) unequal run skills/vitae between the two characters; (b) whether the
comparison retail char had buffs. Diagnosis plan: log both chars' run
skill + vitae + computed runRate; capture our GetRunRate inputs
(effective skill after EnchantSkill, burden) at the moment of comparison;
verify ACE's echoed ForwardSpeed matches what we apply locally
(ApplyServerRunRate) and what retail applies for the same skill. Suspect
list: effective-skill vitae application on the LOCAL run path,
run-rate-cap difference, or an uneven comparison.
**Root cause:** retail `MovementSystem::GetRunRate` (0x006b0950) returns
18/4 = 4.5 ONLY when runSkill == 800 EXACTLY (byte-decoded `fcom [800f];
test ah, 0x44; jp` — the C2/C3 parity equality idiom; <, >, and unordered
all take the general formula). ACE misread the same x87 mush as
`>= 800` ("max run speed?") and our P1 port inherited it via the
ACE cross-reference. Every maxed character therefore ran a flat 4.5
(retail-true ~3.70) — ~21% too fast and completely vitae-independent,
because both the vitae-reduced and unreduced skill sat above 800. The
controlled comparison (33%-vitae +Acdream vs 5%-vitae +Je, both maxed)
showed acdream faster while retail runs them within ~0.4% — exactly the
formula's prediction. The vitae/enchantment chain itself was verified
intact end-to-end via [stat-chain] live capture (vitae 0.67 installed →
eff run 10200 → applied to controller).
---
**Fix:** `==` restores the general formula for all non-800 skills;
golden tests pin 799/800/801 straddle + the maxed-skill vitae
differential; `docs/research/2026-07-30-stat-coupled-movement-pseudocode.md`
§6 corrected with the full byte decode and an explicit "do not re-import
ACE's >= reading" warning.
## #265 — Steep-slope response set: uphill-jump bounce, roof slides lost, edge wedge (TS-4 removal fallout — REVERTED)

View file

@ -204,7 +204,7 @@ AP-127):
`GetRunRate` (0x006b0950) and the `arg3!=0` (PK) branch of `JumpStaminaCost`
(0x006b0a40) have their GENERAL-CASE arithmetic entirely dropped by BN (only
the `EncumbranceSystem::LoadMod`/`800`-skill-cap calls and the `arg3==0`
the `EncumbranceSystem::LoadMod`/`800`-skill-compare calls and the `arg3==0`
ceil expression survive uncollapsed — the same information-loss class as the
x87 mush, just total rather than partial). **ACE is the cross-reference
tiebreaker for those two spots** (`references/ACE/Source/ACE.Server/Physics/
@ -213,13 +213,26 @@ citation style (`PlayerWeenie.cs`'s pre-P1 doc comments already said
"decompiled + ACE MovementSystem" for these two formulas — nothing new here,
just now with a named-decomp address alongside):
- `GetRunRate(load, skill, scaling) = skill>=800 ? 18/4 : ((LoadMod(load) * (skill/(skill+200)*11) + 4) / scaling) / 4`
matches acdream's pre-existing `PlayerWeenie.GetRunRate` exactly (which
hardcoded `scaling=1`); the true retail signature carries a 3rd
`scaling` arg (confirmed by the decomp's own function signature), and
every known call site (`InqMaxRunRate`, `InqRunRate`) passes `1f` — so
porting the full signature is free (no behavior change), just closer to
the retail surface for future callers.
- `GetRunRate(load, skill, scaling) = skill==800 ? 18/4 : ((LoadMod(load) * (skill/(skill+200)*11) + 4) / scaling) / 4`
**§12c correction (#266, 2026-07-30): the 800 branch is EXACT EQUALITY,
not `>=`.** Raw byte decode of 0x006b0950 (PDB-paired binary):
`fild skill; fcom [0x00803b94 = 800f]; fnstsw ax; test ah, 0x44; jp
0x6b097f` — the C2/C3 parity idiom in which `jp` (general path) fires
for `<`, `>`, AND unordered; the `fld [18f]; fdiv [4f]; ret` fall-through
executes only when C3=1/C2=0, i.e. skill == 800 exactly. The general
path decodes instruction-by-instruction to
`(LoadMod(load) * (skill/(skill+200)*11) + 4) / scaling / 4`
(constants 200f @0x00803b8c, 11f @0x00803b88, 4f @0x007c6174, /scaling
from `[esp+0xc]`, final /4f @0x00803b80). **ACE's `>= 800` "max run
speed?" reading is a misread of the same mush and must not be used as a
tiebreaker here** — it flat-lined every maxed character at 4.5 (retail
general formula gives ~3.70) and erased the vitae speed differential
(#266: 33%-vitae +Acdream visibly outran 5%-vitae +Je in acdream while
retail runs them within ~0.4%). `InqMaxRunRate`'s skill=9999 probe gets
the general formula (~3.6961), not 4.5. The true retail signature
carries a 3rd `scaling` arg (confirmed by the decomp's own function
signature), and every known call site (`InqMaxRunRate`, `InqRunRate`)
passes `1f`.
- `GetJumpHeight(load, skill, extent, scaling)` — BN's extent-clamp
micro-branch (pc 006b09b0-006b09ca) is the SAME x87-mush pattern as §3;
ACE's `Math.Clamp(extent, 0, 1)` is the tiebreaker (matches the EXISTING

View file

@ -25,13 +25,28 @@ namespace AcDream.Core.Physics;
public static class MovementSystem
{
/// <summary>
/// <c>MovementSystem::GetRunRate</c> 0x006b0950. Retail-verified 800-skill
/// cap (<c>InqMaxRunRate</c> passes skill=9999 to reach this cap); general
/// case ACE-cross-referenced (BN dropped the arithmetic, see class doc).
/// <c>MovementSystem::GetRunRate</c> 0x006b0950. The 800 branch is an
/// EXACT-EQUALITY sentinel, not a cap: raw byte decode of the PDB-paired
/// binary (#266, 2026-07-30) shows <c>fild skill; fcom [800f]; fnstsw;
/// test ah, 0x44; jp general</c> — the C2/C3 parity idiom that falls
/// through to <c>18/4</c> ONLY when skill == 800 (&lt;, &gt;, and
/// unordered all take the general path). <c>InqMaxRunRate</c> (0x00591b20)
/// passes skill=9999 and therefore gets the general formula's ~3.6961,
/// NOT 4.5. The general-path arithmetic is byte-verified
/// instruction-by-instruction: <c>(loadMod * (skill/(skill+200) * 11) +
/// 4) / scaling / 4</c>.
/// </summary>
/// <remarks>
/// ACE's <c>MovementSystem.GetRunRate</c> reads this branch as
/// <c>&gt;= 800</c> ("max run speed?") — that is ACE's misread of the
/// same x87 mush, NOT a tiebreaker. Porting <c>&gt;=</c> made every
/// maxed character run a flat 4.5 (vs retail's ~3.70) and erased the
/// vitae speed differential entirely (#266). Do not "fix" this back
/// from ACE.
/// </remarks>
public static float GetRunRate(float burden, int runSkill, float scaling = 1f)
{
if (runSkill >= 800)
if (runSkill == 800)
return 18f / 4f;
float loadMod = EncumbranceSystem.LoadMod(burden);

View file

@ -200,7 +200,8 @@ public sealed class PlayerWeenie : IWeenieObject
/// <summary>
/// RunRate = (burdenMod * (runSkill / (runSkill + 200)) * 11 + 4) / 4.
/// Capped at 4.5 when runSkill >= 800. Thin forwarder to
/// Returns 4.5 only for the retail skill == 800 sentinel (byte-decoded
/// exact-equality branch, #266 — NOT a cap). Thin forwarder to
/// <see cref="MovementSystem.GetRunRate"/> — kept for source
/// compatibility with existing golden-value tests.
/// </summary>

View file

@ -1,3 +1,4 @@
using System;
using AcDream.Core.Physics;
using Xunit;
@ -11,10 +12,29 @@ namespace AcDream.Core.Tests.Physics;
public class MovementSystemTests
{
[Fact]
public void GetRunRate_Skill800Cap_Returns4Point5()
public void GetRunRate_Skill800Sentinel_IsExactEqualityOnly()
{
// Retail 0x006b0950 byte decode (#266): fcom [800f]; test ah, 0x44;
// jp — the 18/4 return executes ONLY when skill == 800 exactly.
// 799 and 801 take the general formula on either side of it.
Assert.Equal(4.5f, MovementSystem.GetRunRate(0f, 800), precision: 5);
Assert.Equal(4.5f, MovementSystem.GetRunRate(0f, 999999), precision: 5);
Assert.Equal(3.1994f, MovementSystem.GetRunRate(0f, 799), precision: 3);
Assert.Equal(3.2005f, MovementSystem.GetRunRate(0f, 801), precision: 3);
}
[Fact]
public void GetRunRate_MaxedSkills_UseGeneralFormula_VitaeDifferentiates()
{
// #266 regression: +Acdream at 33% vitae (eff 10200) vs +Je at 5%
// vitae (eff 14463) — retail runs them within ~0.4% of each other
// (~3.70 vs ~3.71), NOT at a flat 4.5. The old ">= 800 cap"
// misread made both saturate identically and erased vitae entirely.
// Tolerance asserts (not precision:) — 3.7125 straddles a 3-decimal
// rounding boundary, the same xunit footgun as the AP-7 case.
Assert.True(MathF.Abs(MovementSystem.GetRunRate(0f, 10200) - 3.6971f) < 1e-3f);
Assert.True(MathF.Abs(MovementSystem.GetRunRate(0f, 14463) - 3.7125f) < 1e-3f);
// InqMaxRunRate's skill=9999 probe: general formula, not 4.5.
Assert.True(MathF.Abs(MovementSystem.GetRunRate(0f, 9999) - 3.6961f) < 1e-3f);
}
[Fact]