Dated source reference. Current release status: handoff state. Full original relative paths are in the developer archive.
College rules feasibility audit — October 13, 2026
Historical pre-implementation assessment. The originalOctober13 probabilities/gaps below are retained as research history, not current completion status. Agreed college mechanics are now implemented in development36/profile4; latest naturalOT/export and18both-endreleasecases passed. Rare live/production-smoke checks are deferred. See current implementation, internal state and verification.
Research only, against OpenHoops development35 / native35a. No game, adapter, rules or package changes, and no new live tests. The user requested an evidence-based assessment of probable versus impractical college-rule changes before implementation. Current goal remains a Windows-first extension of NBA 2K14, not full engine reconstruction.
Verdict and review boundaries
No requested core college rule is established to be impossible. A substantial college conversion is probable through targeted engine hooks/replacements. That is an engineering judgment, not a demonstrated NCAA mode. Possession-arrow administration, one-and-one execution, AI adaptation and complete officiating parity have the largest gaps. Do not promise full rulebook fidelity, stock presentation accuracy, arbitrary mod coexistence or save/resume support at this stage.
Reviewed the archived tool/community catalogs, Elan source and README, Sigil correction, Blaze SDK/package notes, ShotMeter/GreenMeter research, local rule/clock/geometry/FT/timeout findings, selected original Ghidra routines, retained verification results, and current adapter/client assumptions. Searched the indexed library for halves, one-and-one, foul-out and possession-arrow references. This is an assessment of the available archive and existing code recovery, not a fresh exhaustive disassembly of every downloaded binary. Uninspected ShotClockReset.dll internals remain a lead, not evidence of a working NCAA reset policy.
Evidence manifest: college-rules-audit/manifest.json. Existing tests are reviewed, not rerun or reclassified as new live proof. Third-party sources retain their original attribution and version/hash manifests.
Feature assessment
| Feature | Judgment | Concrete evidence | Unresolved work / acceptance boundary |
|---|---|---|---|
| Disable defensive three seconds and NBA back-to-basket restriction | High confidence | Existing original-handler rule switches, API readback and original-menu corroboration in dev32 | New NCAA preset/application timing and actual gameplay enforcement regression; menu/readback proof alone does not test every violation |
| Ten-second backcourt | High confidence for core violation | Actual timer selector table; 216 captured-x86 dispatch/handler cases; controlled selector/timer fixture emitted kind1 and retained 10-SEC VIOLATION UI | Live fixture accelerated remaining timer; not natural ten-second timing. NCAA count starts/resets and 2026 display precision still require tests |
| Thirty-second full shot clock | High confidence | Configured reset parameter separated from remaining-time clock; reset writers and original timer operation recovered; Elan source provides historical precedent | Apply before initialization; cover near-period-end behavior, reset causes and AI pacing |
| Twenty-second offensive-rebound/frontcourt reset policy | Probable | Reset writers, conditional-minimum operation, committed rebounds/corrections and shot-clock observations mapped | Fourteen-second minimum in original game is NOT its OR rule: stock observed OR reset was24. Intercept actual reset cause; rebound-counter polling is too late/ambiguous |
| Two halves, ten/twenty minutes | Probable, substantial integration | Period setup/reset/OT routines, period-duration parameter, end-game lifecycle and exact Elan period2-to4 source | Original setup explicitly treats period3 as halftime and period5+ as OT. Need correct direction switch, bonus resets, finalization, CPU substitutions, presentation and stat-period mapping. Current pregame minutes1..12 is an API/menu limit, not proof of an engine ceiling |
| Five-minute OT independent of half duration | High confidence in a targeted path, unimplemented | Original882800 duration routine bounded-tested; natural double OT observed | Original duration scales from quarter minutes; a20-minute half cannot inherit its current scaling blindly. Test repeated OT and tied-final-shot/FT boundaries |
| Correct college three-point geometry and scoring | Probable; stronger than score-value hacks | Actual244-byte89FF10 four-point classifier; separate radius/corner-width/transition; both directions and1,266 captured-code comparisons; caller sets scoring flag400 | Exact college dimensions, rendered court alignment, body-point extraction, boundary shots, all FT paths, shot chart and AI spacing must agree. Alternate bank resembles international dimensions but has no verified league identity |
| Twelve-foot lane with actual lane enforcement | Plausible, lower confidence than arc | Court asset tools plus offensive-three-second timer/caller leads | Lane predicate dimensions, free-throw spots and player/AI placement not recovered as an integrated custom geometry feature |
| Five-foul disqualification | Probable, moderate confidence | PF writer/counters, roster actors and existing engine substitution/disqualification behavior provide a focused target | No verified foul-out threshold/replacement hook yet. Avoid falsifying PF stats to trick the sixth-foul path. Must handle CPU rotation, technical accounting and exhausted bench |
| Made-basket clock policy | Probable, moderate confidence | Made-shot paths, timer start/stop operations and period context known | Must isolate make-specific timing decision; generic stop routine is invoked thousands of times, including FTs. Need first-half and second-half/OT boundary cases without suppressing whistle stops |
| Four team timeouts with NCAA carryover | Probable | Counts/types/countdown/phase observed; original8C7EF0 entry passes96 bounded cases; live normal timeout captured | Allowance initialization/late-period trimming/carryover, UI, AI and short-timeout duration differ. Type shown during timeout is not necessarily allowance pool charged |
| TV timeouts | Plausible, meaningful new subsystem | Timeout entry and resume context identified; phase17 corroborated live | Authoritative eligible-dead-ball boundary, queue/merge rules, no team charge, consecutive-break handling, interrupted FT/inbound resume and fatigue effects. Mocked eligibility in timeout tests does not establish legal scheduling |
| NCAA bonus thresholds and OT foul carryover | Probable | Qualifying-counter writer, threshold predicates, reset policy and real bonus award consumer recovered; natural default-rule bonus award observed | NCAA qualifying-foul accounting,7/10 thresholds, late-period NBA exception removal, OT carryover and UI must change together. Nondefault selectors are not proven NCAA presets |
| One-and-one | Plausible but one of the highest-risk core items | FT award construction, type/total/remaining, make/miss paths and natural two-shot sequence observed | No confirmed native one-and-one state. Need conditional second attempt plus live first-miss rebounding, lane lineup, substitutions, horn timing and stats. Existing unknown FT kind3 must not be named one-and-one without evidence |
| Alternating-possession arrow | Plausible but weakest core evidence | Team context, opening/OT tips, inbound and stop paths provide ingredients | No verified arrow state/consumer. Likely new policy: establish arrow, held-ball alternatives, halftime allocation, throw-in completion/violation/foul exceptions and reset on each OT tip. Physical attachment is not legal team control |
| Remove NBA timeout advancement | Plausible, moderate confidence | Timeout and inbound placement path8BCA60 identified | That path also handles direction/late-clock logic. Need specific advancement predicate, not unconditional inbound relocation |
| Closely guarded holding violation | Plausible, low-to-moderate confidence | Game asset vocabulary includes closely guarded labels; actor positions/holding and timer machinery exist | Vocabulary is not a mapped detector. No proven six-foot holding-only rule or complete caller map |
| Exact NCAA traveling, continuation, charging, technical/flagrant and challenge policy | Low confidence as a complete fidelity promise | Some foul handlers, classification, FT/scoring and shot-action leads | Contact/animation/gather detection, severity categories, replay/review state and distinct penalties are largely unmapped. Full parity is a separate major project |
Why the existing research is useful
Our game-thread adapter already executes guarded original setters and publishes coherent snapshots/ordered records. We can target authoritative decisions instead of correcting a score or timer after an external poll. The35a fixture and previous natural OT/bonus fixtures establish a working platform for further experiments; they do not validate new college rules.
The geometry finding is especially concrete. analysis/live-data/three-point-verification.json compares the exact predicate with a float32 transcription across1,266 cases. analysis/rule-completion/ghidra/008bdeb0.c additionally shows the shooting award path either calling89FF10 for the current owner or reading classification flag400 for a detached ball. This is a credible connection between court geometry and two/three-shot awards, but inferred Ghidra output and bounded branches do not prove all live callers or correct AI spacing.
Two-halves support must include our API. api/openhoops_api/presentation.py explicitly labels halftime from rawperiod3 and advertises four_quarter_exhibition. protocol.py exports four regulation score buckets; scoring.py treats rawperiod>4 as overtime. We can translate engine periods into logical halves or replace more lifecycle code, but choosing that design requires proof. Skipping raw periods without adapting these contracts would corrupt labels/reconciliation.
What the community archive contributes
- Elan 2K Mod, elanjcoop, archived commitf77f0faeee55a1c077f52c79d2b4e8e5db0d47f6: real readable experiments with period length, halves, clocks and scoring.
main.pyreplaces remaining8 with10, changes rawperiod2 to4 and reacts to OR counters. README admits shortened-three shooting-foul and reset failures. Useful leads, not a complete college mode. No code incorporated during this audit. - Blaze0.5.1 / ShotMeter1.8.6 / GreenMeter1.7.0: native patch/plugin and shot-action precedents. Blaze SDK explicitly runs lifecycle callbacks on its worker, not the game thread. Existing OpenHoops game-thread work is more directly useful for our rule transactions. Package claims do not prove compatibility with our adapter or NCAA semantics.
- NBA2K14 HOOK2.0.3: four-point Lua strings credit Elan; ShotClockReset.dll exists. Compiled chunks/file names do not demonstrate different or more complete rule enforcement.
- RED MC5.0, Sigil1.6.3, Roster Lab1.7.0: roster/team/asset preparation and schemas. Sigil explicitly withdrew its mistaken Era rules panel after recognizing controller settings. No verified college-rule preset follows from that panel.
- IFF/Blender/court/scoreboard tools, BodyHook1.0/1.1 and audio tools: support college visual identity, asset workflows and presentation. They do not settle rule decisions. College recruiting/ranking tools are external management examples.
- Old Cheat Engine0.5 table warns some slider changes were only visual. We must retain that distinction between displayed values and simulation behavior.
See Elan, tool findings, community hooks, Sigil, live-data geometry, rule policy, newer live evidence.
Claims to reject and next decision
Reject: a hidden proven complete NCAA mode; geometry by texture replacement alone; semantics from selector names; reliable rules from HTTP polling of clock/stat deltas; automatic support for all EXEs/mod stacks; unchanged four-quarter API/result schemas working as a two-half contract; guaranteed save/resume without new serialized college state. None is supported by the archive.
No technical impossibility has been established for the core list. For scope planning, do not commit now to perfectly NCAA-specific AI, every contact/technical/challenge edge case, or stock commentary/cinematics that accurately reflect all new rules. These are high-cost/weak-evidence areas, not mathematical impossibilities.
Recommended future proof order (not executed): (1) a true two-half match through halftime, finish and repeated OT with correct API periods; (2) one-and-one first make/miss plus7/10 bonus and OT carryover; (3) arrow/held-ball and inbound administration; (4) geometry plus scoring/FT alignment; (5) timer/timeout policies and AI/presentation regressions. Failure on one-and-one or arrow determines what must be selectively replaced; it does not by itself justify full engine decompilation.