# What still exceeds five?

Follow-up: [The mamenchisaurid locality review](mamenchisaurid-locality-review.html) supersedes the two broad Daye/Chaya target comparisons below with explicit Tibet/Xizang clues. Their province-only maxima are now two. The original selected-clue counts are retained as a checkpoint.

Current rule: allow five indistinguishable remaining candidates; free referenced support starts at six, when neither the tree nor standard clues can distinguish them.

The after-locality comparison assumes all three provisional standard clues are revealed, the chosen locality is available and every preceding guess observation is applied.

All 484 selected locality comparisons across 63 source-clue pools were assessed (21,656 reaching-guess contexts).

- 473 targets are already within five after their proposed locality and reaching guess.
- 11 targets have larger states but an answer-independent sequence of informative guesses brings their actual feedback route to five or fewer.
- 0 supported clue targets still end in a blind crowd above five under that sequence.

**This does not make every large branch playable without a locality clue.** The standard-clue baseline still contains larger crowds. The check asks whether an additional, more specific clue is needed after the proposed locality and useful navigation.

## Every target needing navigation under the new limit

| Target | Chosen locality | Largest initial crowd | Largest terminal crowd for this target |
| --- | --- | ---: | ---: |
| Bisticeratops | Kirtland Formation · US | 6 | 5 |
| Navajoceratops | Kirtland Formation · US | 6 | 5 |
| Terminocavus | Kirtland Formation · US | 6 | 5 |
| Yeneen | Bajo de la Carpa Formation · AR | 7 | 5 |
| Daliansaurus | Yixian Formation · CN | 6 | 5 |
| Sinusonasus | Yixian Formation · CN | 6 | 5 |
| Wendiceratops | Oldman Formation · CA | 6 | 5 |
| Damalasaurus | Daye Group · CN | 34 | 5 |
| Microdontosaurus | Chaya Group · CN | 34 | 5 |
| Datousaurus | Shaximiao Formation · CN | 10 | 5 |
| Shunosaurus | Shaximiao Formation · CN | 10 | 5 |

All other targets have validated maxima of five or fewer. The full JSON preserves all 484 outcomes, all greater-than-five target routes and all baseline pools.

## The sauropod distinction

The earlier country-only Daye Group comparison for Damalasaurus and Chaya Group comparison for Microdontosaurus can initially retain 34 candidates. An Abrosaurus probe is selected from the remaining state alone. For either actual target, its Sauropoda feedback leaves at most five basal sauropods. Those five are acceptable under the new rule; their incomplete scientific profiles still need work.

A different possible response to the same probe would leave 19 mamenchisaurids. In these country-only Daye/Chaya comparisons, all 19 survive because relevant locality metadata is unknown; none is a recorded Daye match, and neither supported clue target lies in that branch. This is a footprint-completeness warning, not evidence for a 19-genus Daye assemblage or proof that those 19 genera fail with their own supported locality clues.

### Possible 19-candidate Mamenchisauridae outcome

Daanosaurus, Eomamenchisaurus, Tonganosaurus, Xinjiangtitan, Huangshanlong, Qijianglong, Anhuilong, Omeisaurus, Tienshanosaurus, Analong, Rhomaleopakhus, Jingiella, Tongnanlong, Bellusaurus, Hudiesaurus, Zigongosaurus, Klamelisaurus, Chuanjiesaurus, Yuanmousaurus.

For these tested contexts: Damalasaurus, Microdontosaurus.

Known locality matches: none; all retained candidates have relevant unknown locality metadata.

Decision: retain the possibilities conservatively; complete their locality footprints and standard facts. Do not add a mandatory discovery-date clue or approve these as Daye hidden answers on this hypothetical outcome alone.

## Crowds before locality support

The representative baseline identifies 46 of 63 matching standard-clue pools needing further evidence even with a five-candidate limit. A useful locality clue can supply that evidence. Large pool counts alone are not failures after locality support.

| Source parent | Standard clues | Complete provisional siblings | Largest representative crowd before locality |
| --- | --- | ---: | ---: |
| Hadrosauroidea | herbivore · Cretaceous · Asia | 24 | 44 |
| Nodosauridae | herbivore · Cretaceous · Americas | 20 | 20 |
| Mamenchisauridae | herbivore · Jurassic · Asia | 19 | 20 |
| Chasmosaurinae | herbivore · Cretaceous · Americas | 16 | 20 |
| Euhelopodidae | herbivore · Cretaceous · Asia | 15 | 14 |
| Titanosauria | herbivore · Cretaceous · Americas | 12 | 64 |
| Heyuanninae | omnivore · Cretaceous · Asia | 11 | 10 |
| Parvicursorinae | carnivore · Cretaceous · Asia | 11 | 10 |
| Rebbachisauridae | herbivore · Cretaceous · Americas | 11 | 15 |
| Rinconsauria | herbivore · Cretaceous · Americas | 11 | 12 |
| Elasmaria | herbivore · Cretaceous · Americas | 10 | 10 |
| Ankylosaurinae | herbivore · Cretaceous · Asia | 9 | 10 |
| Neoceratopsia | herbivore · Cretaceous · Asia | 9 | 18 |
| Somphospondyli | herbivore · Cretaceous · Asia | 9 | 53 |
| Therizinosauridae | herbivore · Cretaceous · Asia | 9 | 10 |
| Troodontidae | carnivore, omnivore · Cretaceous · Asia | 9 | 22 |
| Elasmaria | herbivore · Cretaceous · Oceania | 8 | 8 |
| Leptoceratopsidae | herbivore · Cretaceous · Americas | 8 | 8 |
| Lithostrotia | herbivore · Cretaceous · Americas | 8 | 24 |
| Somphospondyli | herbivore · Cretaceous · Americas | 8 | 74 |
| Ankylosauridae | herbivore · Cretaceous · Asia | 7 | 17 |
| Centrosaurinae | herbivore · Cretaceous · Americas | 7 | 20 |
| Furileusauria | carnivore · Cretaceous · Americas | 7 | 9 |
| Hadrosauroidea | herbivore · Cretaceous · Americas | 7 | 50 |
| Kritosaurini | herbivore · Cretaceous · Americas | 7 | 6 |
| Megaraptoridae | carnivore · Cretaceous · Americas | 7 | 7 |
| Microraptorinae | carnivore · Cretaceous · Asia | 7 | 7 |
| Sauropoda | herbivore · Jurassic · Asia | 7 | 43 |
| Titanosauria | herbivore · Cretaceous · Asia | 7 | 28 |
| Brachyrostra | carnivore · Cretaceous · Americas | 6 | 15 |
| Caenagnathinae | omnivore · Cretaceous · Americas | 6 | 6 |
| Carcharodontosauridae | carnivore · Cretaceous · Americas | 6 | 8 |
| Dromaeosauridae | carnivore · Cretaceous · Asia | 6 | 24 |
| Eusauropoda | herbivore · Jurassic · Asia | 6 | 33 |
| Eutitanosauria | herbivore · Cretaceous · Americas | 6 | 6 |
| Lognkosauria | herbivore · Cretaceous · Americas | 6 | 6 |
| Ornithomimidae | herbivore, omnivore · Cretaceous · Americas | 6 | 7 |
| Ornithomimidae | herbivore, omnivore · Cretaceous · Asia | 6 | 8 |
| Tyrannosauroidea | carnivore · Cretaceous · Americas | 6 | 19 |
| Tyrannosauroidea | carnivore · Cretaceous · Asia | 6 | 16 |
| Unenlagiinae | carnivore · Cretaceous · Americas | 6 | 6 |
| Arenysaurini | herbivore · Cretaceous · Europe | 5 | 6 |
| Brachylophosaurini | herbivore · Cretaceous · Americas | 5 | 6 |
| Neornithischia | herbivore · Jurassic · Asia | 5 | 16 |
| Nodosauridae | herbivore · Cretaceous · Europe | 5 | 6 |
| Styracosterna | herbivore · Cretaceous · Europe | 5 | 28 |

## Decisions

- Keep the six-plus trigger and five-candidate destination for free support.
- Baruungoyot’s five alvarezsaurids no longer require an extra discovery-date clue solely for crowd size.
- Keep proposed locality evidence for large ordinary-clue pools; allowing five does not remove the need for it.
- No additional discovery-date clue is currently required by a supported target’s count result in this 484-comparison screen. Discovery history remains a possible later fallback.
- Continue the separate standard-fact, specimen-identity and footprint reviews. Passing this count check does not activate expanded hidden answers.

## Methods and limits

- With all three provisional standard clues revealed, all 484 chosen locality comparisons in all 63 matching source-clue pools are assessed. Their wider candidate sets and reaching-guess counts come from the complete locality validation, which also covers 108 large parent groups.
- If a validated maximum is already at most five, further observations cannot increase that state. For every larger comparison, reconstruct all previously recorded five-plus reaching states, use the same all-genus informative-probe selection, and stop at five.
- Every next probe is selected from the whole remaining state, without knowing the answer. The target then supplies the actual relationship feedback to trace its route through that fixed plan.
- Also inspect every possible wrong outcome, including competitors retained only by uncertain locality metadata. Such hypothetical outcomes are reported separately from failures for the supported clue target.
- All 63 direct-parent standard-clue pools receive a before-locality comparison with a representative supported genus. A largest reaching state demonstrates whether that pool still needs a locality clue or other evidence; it is not an exhaustive playthrough for every baseline target.
- The older four-candidate assessment is retained as a historical checkpoint. Current gameplay support starts at six remaining blind candidates and should leave at most five.
- Passing the candidate-count test does not approve a hidden answer, a complete fossil footprint or the future Expert working tree. Period, region, diet and historical taxon issues remain separate.
- Locality nonmatches are catalogue diagnostics, not proof of scientific absence. Known non-primary occurrences, qualified referrals and relevant unknown metadata remain possible.
- This is a reproducible informative navigation strategy, not an exhaustive optimal game policy or a guarantee that players select the same guesses.
- The frozen PBDB source hierarchy can contain unresolved sibling groups that a reviewed working tree may later distinguish. No new branches or roster exclusions are introduced.

Reproduce with `python evaluate-five-limit.py`. Historical five-plus/four-candidate reports remain unchanged; this report supersedes their gameplay threshold.