# Chinn 1200 BID AUDIT — v2 (UNIFIED)

*Triangulated from 3 independent sources: Atom Library v2 · Spec Manifest · Bluebeam Inspector*

**Audit run:** 2026-04-30 17:38 UTC
**Sources:**
- Atom Library v2: 76 closed jobs · 18 cohorts
- Spec Manifest: 3 outline + supplemental specs parsed
- Bluebeam Inspector: 34 markup links · 1 stale Filesets · 13 missing canonical tags

## Headline

**Bid total:** $4,700,000 · **215u** · **1,648 permit fx** · **12% O&P**
**Per-unit:** $21,860/u · **Per-fixture:** $2,852/fx

**Findings:** 🚨 1 CRITICAL · ⚠ 2 HIGH · · 3 MEDIUM · ✓ 1 CONFIRMED-OK
**Estimated $ impact of action items:** $60,000

## Triangulated Findings

Each finding is corroborated by 1-3 independent sources. Multi-source findings are HIGH confidence; single-source findings are flagged for manual review.

### 🚨 CRITICAL — Hot-water submeter system priced at $0 — confirmed scope omission
**Corroborating sources:** `spec` · `bluebeam`
**Estimated $ impact:** $60,000

**Evidence:**
- **spec:** Spec 010000 Outline Specifications (Plumbing section, p3): "Provide water meters for hot water supply only in units with remote readers. Install in closet. Include transceivers and repeaters."
- **bid:** Bid sheet ESTIMATE tab rows 64-66: 3 submeter-related rows with qty=0 (cold meters $50/ea, hot meters/transceivers $120/ea, repeaters $12k/set — all priced but qty=0)
- **bluebeam:** Canonical markup template subjects WATER METER, TRANSCEIVER, REPEATER NOT present in bid sheet's Bluebeam template (still cloned from 2017 Fairfield Bear Creek)
- **atom_calibration:** No directly comparable atom (n=0 closed jobs with hot-only submeter scope at this scale). Estimate from line-item pricing: 215 units × $120 transceiver + ~$60k repeaters/setup ≈ $50-90k

**Action:** Add WATER METER + TRANSCEIVER + REPEATER markup tags to Bluebeam template, re-quantity from Plans, price scope at calibrated $50-70k. Confirm with Chinn whether owner is taking this scope (unlikely given spec language).

### ⚠ HIGH — Bluebeam template Fileset references 8-year-old project plans
**Corroborating sources:** `bluebeam`

**Evidence:**
- **bluebeam:** BbQuantityLink XML in bid sheet still references Fileset 'Fairfield Bear Creek April 2017'. PDF paths: \\server\management\Estimating\Fairfield - Bear Creek\Plans- First City Submittal- Bid Set\... (2017 plans).
- **implication:** Bluebeam markup quantities may be carrying over from the cloned template instead of being re-counted from Chinn 1200's actual Plans.pdf. Quantities should be verified by walking the plans manually OR by re-pointing the Fileset to /Chinn 1200/Plans.pdf and re-running Bluebeam count.

**Action:** Re-point Fileset GUID 2b847260-9256-4658-b052-cd77d893cdca to current Chinn 1200 Plans.pdf path. Verify each canonical subject's count against the new plans. Save the bid sheet with updated XML.

### · MEDIUM — Bid $/unit $21,860 sits +58% above Chinn 1200 archetype blend
**Corroborating sources:** `atom`

**Evidence:**
- **atom_blend:** 4-cohort weighted blend (gc_chinn 57% + chinn1200_archetype 29% + size_large 14%): median $13,813/u, P25-P75 $9,332-$17,688
- **calculation:** This bid: 215u × $21,860/u = $4,700,000. Archetype blend expectation: $2,969,761.
- **bid:** Excess vs archetype: $+1,730,239
- **interpretation:** Justified by heat-pump central plant (no archetype comp at this scale), Bellevue downtown urban site, SD-set bid scope risk. Worth verifying scope premium attribution against Phase 1 audit's variance attribution.

**Action:** Cross-reference variance with Phase 1 audit's attribution table. Residual unattributed premium = strategic margin posture for SD-set risk.

### · MEDIUM — Heat-pump central plant atom is single-comp (n=1) — calibration uncertainty
**Corroborating sources:** `atom`

**Evidence:**
- **atom_calibration:** Heat-pump central plant cohort has n=1 closed comp(s) (KOZ #2099, Trane vendor at $290k equip). NYLE 4-pack at $448k is +54% above the only comp.
- **spec:** Spec 010000 confirms 'Water Heating: Via heat pumps' but doesn't specify NYLE vs Trane.
- **bid:** $555k total heat-pump plant (4 NYLE H400 × $112k + 5 storage tanks × $19k + circ + controls)
- **implication:** NYLE pricing is +54% above KOZ Trane. Either NYLE genuinely costs more, OR the Trane comp from KOZ understates current heat-pump pricing (KOZ closed 2024). Get a competing Trane quote to validate.

**Action:** Request a Trane HPWH quote for the same scope as alternative. If Trane comes in at $290-350k vs NYLE $448k, switch to Trane and capture $100-150k. If Trane confirms $400k+ market price, NYLE is in-line and the comp atom needs updating.

### ✓ LOW — 2 scope items from outline specs confirmed present in bid
**Corroborating sources:** `spec` · `bid`

**Evidence:**
- **items:**
  - {'scope_item': 'Heat-pump water heating central plant', 'expected_range': [290000, 600000], 'matches_in_bid': 3}
  - {'scope_item': 'Dog wash stations (2 elevated SS grooming tubs with hot/cold)', 'expected_range': [15000, 20000], 'matches_in_bid': 4}

**Action:** No action — these scope items are correctly priced.

### · MEDIUM — Bluebeam template missing 6 specialty markup tags
**Corroborating sources:** `bluebeam`

**Evidence:**
- **bluebeam:** Canonical specialty subjects not in template: DOG WASH, MOP SINK, FROST LOCKER, EYE WASH, DRINKING FOUNTAIN, SAUNA
- **implication:** Specialty fixtures are priced manually in the bid (DOG WASH $16k, SAUNA $6k, etc.) without Bluebeam quantity verification. If quantity changes mid-project, the manual line item won't auto-update.

**Action:** Add specialty subjects to canonical Bluebeam template. For Chinn 1200: mark up 2 dog wash tubs, sauna, frost locker locations on Plans.pdf.

### ⚠ HIGH — Cast iron stack quantities not pulled from Bluebeam markup
**Corroborating sources:** `bluebeam` · `bid`

**Evidence:**
- **bluebeam:** CAST IRON STACK markup tag NOT in bid sheet's Bluebeam template (despite being in canonical template)
- **bid:** Cast iron stack lines in bid: $215,208 (manually entered: 768.6 LF × $120/$160/$25 across multiple lines)
- **implication:** Cast iron stack count is one of the largest line items at ~$280k+ but quantity is hand-entered, not verified against plans. Risk of under/over-count is real.

**Action:** Add CAST IRON STACK markup tag, walk plans to count stacks per floor (8 above + 3 below = 11 levels), verify 768.6 LF total or update.

## Atom Library v2 calibration

Chinn 1200 is calibrated against a **4-cohort weighted blend** (`gc_chinn` 57% · `chinn1200_archetype` 29% · `size_large` 14%; `heat_pump_central` dropped at n=1 < blend threshold).

| Atom | Blend Median | P25–P75 | This Bid | Variance |
|---|---:|---:|---:|---:|
| `ORIG_BID_USD_PER_U` | $13,813/u | $9,332–$17,688 | $21,860/u | +58% |
| `TOTAL_DIRECT_COST_USD_PER_U` | $8,002/u | $5,234–$10,594 | $19,518/u | +144% |
| `TOTAL_LABOR_USD_PER_FX` | $414/fx | $293–$540 | $778/fx | +88% |
| `TOTAL_MATERIAL_USD_PER_FX` | $427/fx | $395–$564 | $1,204/fx | +182% |

## Spec manifest coverage

Outline specifications confirm scope items in bid. Missing items = scope omissions.

| Status | Spec scope item | Expected $ | Found in bid? |
|---|---|---:|---|
| 🚨 | Hot-water submeters with transceivers + repeaters | $50,000–$70,000 | 🚨 NO ($0 priced) |
| ✓ | Heat-pump water heating central plant | $290,000–$600,000 | ✓ $555k priced |
| ✓ | Dog wash stations (2 elevated SS grooming tubs with hot/cold) | $15,000–$20,000 | ✓ $16k priced |
| ✓ | No gas to amenity spaces | — | ✓ Confirmed (no gas pricing) |
| ✓ | Sauna + frost locker rough-in (AmFinn vendor-supplied) | $5,000–$10,000 | ✓ $6k rough-in |

## Bluebeam template health

- **Total markup links:** 34
- **Filesets:** 1
- **Stale Filesets:** 1
  - 🚨 `Fairfield Bear Creek April 2017` (linking to old project plans)
- **Canonical subjects missing:** 13
  - drainage: UNDERSLAB
  - equipment: SP-3
  - piping: RECIRC HW MAIN
  - riser: CAST IRON STACK
  - specialty: DOG WASH, SAUNA, FROST LOCKER, DRINKING FOUNTAIN, MOP SINK, EYE WASH
  - submeter: WATER METER, TRANSCEIVER, REPEATER

## Profitability forecast

OWP closed-portfolio gross margin: median 38.4% · P25 31.1% · P75 45.4% (n=76).

At bid: direct cost $4,196,269 + O&P $503,731 = $4,700,000.

**Range scenarios (gross margin × bid):**
- Conservative (KOZ-anchored heat-pump comp): 32.4% gross margin → **$1.52M profit**
- Likely (4-cohort blend median): 37.6% gross margin → **$1.77M profit**
- Optimistic (NG-R Roosevelt anchor): 47.6% gross margin → **$2.24M profit**

## Prioritized action items (pre-2nd-bid)

1. 🚨 **Hot-water submeter system priced at $0 — confirmed scope omission** — Add WATER METER + TRANSCEIVER + REPEATER markup tags to Bluebeam template, re-quantity from Plans, price scope at calibrated $50-70k. Confirm with Chinn whether owner is taking this scope (unlikely given spec language).
2. ⚠ **Bluebeam template Fileset references 8-year-old project plans** — Re-point Fileset GUID 2b847260-9256-4658-b052-cd77d893cdca to current Chinn 1200 Plans.pdf path. Verify each canonical subject's count against the new plans. Save the bid sheet with updated XML.
3. ⚠ **Cast iron stack quantities not pulled from Bluebeam markup** — Add CAST IRON STACK markup tag, walk plans to count stacks per floor (8 above + 3 below = 11 levels), verify 768.6 LF total or update.
4. · **Bid $/unit $21,860 sits +58% above Chinn 1200 archetype blend** — Cross-reference variance with Phase 1 audit's attribution table. Residual unattributed premium = strategic margin posture for SD-set risk.
5. · **Heat-pump central plant atom is single-comp (n=1) — calibration uncertainty** — Request a Trane HPWH quote for the same scope as alternative. If Trane comes in at $290-350k vs NYLE $448k, switch to Trane and capture $100-150k. If Trane confirms $400k+ market price, NYLE is in-line and the comp atom needs updating.
6. · **Bluebeam template missing 6 specialty markup tags** — Add specialty subjects to canonical Bluebeam template. For Chinn 1200: mark up 2 dog wash tubs, sauna, frost locker locations on Plans.pdf.
7. ✓ **2 scope items from outline specs confirmed present in bid** — No action — these scope items are correctly priced.

## Closed-loop calibration hooks

When Chinn 1200 closes (post-construction, after JDR is filed), the following atoms will auto-update:

- `HEAT_PUMP_CENTRAL_PLANT` cohort: n=1 → n=2 (KOZ + Chinn 1200). Variance band tightens.
- `gc_chinn` cohort: n=6 → n=7. Chinn-specific behavior pattern sharpens.
- `bellevue` cohort: n=3 → n=4. Bellevue area pricing reference grows.
- `chinn1200_archetype` cohort: n=12 → n=13.

After closing, re-run `build_atom_library_v2.py` to refresh atom library to v3. Each closed bid sharpens future calibrations.

---

## Provenance

- **Atom Library:** [`ATOM_LIBRARY_v2.json`](../atoms/ATOM_LIBRARY_v2.json) (built 2026-04-30T13:52:17.631881+00:00)
- **Spec Manifest:** [`CHINN_1200_SPEC_MANIFEST_v1.json`](../manifests/CHINN_1200_SPEC_MANIFEST_v1.json)
- **Bluebeam Inspection:** [`CHINN_1200_BLUEBEAM_INSPECTION_v1.json`](../inspections/CHINN_1200_BLUEBEAM_INSPECTION_v1.json)
- **Canonical Markup Template:** [`canonical_markup_template.json`](../canonical_markup_template.json)
- **Audit script:** `audit_bid_v2.py`

*Re-run anytime: `python3 audit_bid_v2.py <bid_sheet.xlsx>`. Audit re-pulls all 3 sources fresh.*