Build Around Observed Contributions
Start with 
Roles overlap. Record both the bottle’s standalone profile and its observed contribution at a stated proportion in a particular blend. Grain, age and producer descriptions guide candidate selection but do not establish performance.
Provisional Roles And Retained Candidates
Oak And Mature-Character Contributors
Test desirable wood, leather, tobacco and tea impressions separately from excessive bitterness and drying astringency. “Age anchor” does not transfer an age claim to the blend.
- Elijah Craig Small Batch.
- Elijah Craig Barrel Proof — separate expression; record batch and actual proof.
- Knob Creek 9-Year Single Barrel Reserve, 120 proof — verify the particular label.
- Knob Creek 12-Year Old.
- Russell’s Reserve Single Barrel Bourbon.
- Wild Turkey Rare Breed Bourbon.
Body And Perceived-Softness Candidates
These wheated/wheat selections are candidates, not proof that wheat guarantees silkiness or removes heat. Test body, coating, aroma and heat separately.
- Ben Holladay Soft Red Wheat Rickhouse Proof.
- Larceny Barrel Proof.
- Maker’s Mark Cask Strength.
- Middle West Spirits Straight Wheat Whiskey, Cask Strength.
The previously specified Maker’s Mark trial bottle is 112 proof; do not apply that value to every release.
Fruit, Roast And Dessert Contributors
Keep fresh fruit, dried/stewed fruit, roast/nut and dessert impressions distinct. Do not assume increased viscosity, chemical reactivity or a “multiplier” effect.
- Chattanooga Whiskey Cask 111.
- Four Roses Private Selection Barrel Proof, OBSV — retain the intended candidate but verify exact label/recipe; do not conflate standard Single Barrel and private barrel-proof expressions.
- Jack Daniel’s Single Barrel Barrel Proof.
- Old Forester 1910.
- Old Elk Double Wheat, Master’s Blend.
- Woodford Reserve Double Oaked.
Spice Contributors
Rate baking spice, pepper and herbal impressions separately. Assess fruit independently; neither rye nor dark fruit is mandatory for balance.
- Jack Daniel’s Single Barrel Barrel Proof Rye.
- Pikesville Straight Rye, 110 proof.
- Sagamore Cask Strength Rye.
- Alberta Premium Cask Strength Rye — exact expression/label to verify; do not assume the original “Straight” wording.
- Wild Turkey Rare Breed Rye.
Finished-Whiskey Accents
- Barrell Vantage — a candidate for differentiated oak character, not a guaranteed rancio effect.
The producer identifies Mizunara, French and toasted American oak finishes. These details describe production, not a tested contribution to our blend. Producer reference.
Optional Brightness And Lift
Assess whether existing candidates supply useful fresh fruit, citrus, floral or herbal contrast. No additional purchase is prescribed. A rich blend need not maximize darkness in every dimension.
Separate Experimental Branch: Non-Whiskey Modifiers
Retain these original candidates separately from whiskey-only formulation:
- Delord Bas Armagnac, 25-Year Old.
- Bodegas Tradición Oloroso VORS, labeled 30 years — exact bottle to verify.
- Château Suduiraut Sauternes — record vintage.
- Graham’s 20-Year-Old Tawny Port.
Direct addition is not equivalent to cask finishing or accelerated aging. These ingredients can change alcohol strength, sweetness, acidity and flavor together; an apparent improvement must not automatically be attributed to mature aromas. Keep whiskey-only trials separate and disclose all ingredients. No commercial labeling determination is made here.
Record Each Component And Trial
Use the existing Whiskey Profiles and Whiskey Blends records rather than creating a competing catalog. The following is a recording specification, not a claim that new database fields have already been implemented:
- Exact expression, batch/barrel, recipe when known, actual proof and source of specifications.
- Standalone observations and proposed roles, labeled “untested.”
- Base/recipe version, proportion, measured amounts, water treatment and estimated or measured final strength with method stated.
- Session date, vessel, rest interval, order, sensory observations and unwanted effects.
- Result versus unchanged base, target-fit rating and repeat-session outcome.
- Evidence status: proposed; observed once; repeated personally; evaluated by additional tasters. These are proposed labels, not existing statuses asserted.
- Cost per trial or finished volume, recorded separately from sensory scoring.
Practical Workflow And Advancement
- Use current stock before buying more. Seek a new component only for a documented gap.
- Begin with a simple base and limited variations; do not assume every role needs its own bottle.
- Test contributions, preserving the unchanged-base comparison. A successful addition may perform differently at another proportion or proof.
- Separate native-proof and normalized-proof sessions; record water additions rather than treating dilution as neutral.
- Advance only after the agreed minimum quality requirements, at least 4 of 5 target fit, and leading-group performance in a separately prepared repeat session. Use the canonical version 0.3 house scoring rubric in
Default is one person doing preparation and evaluation. Keep recipes concealed where practical and randomize order, but describe the remaining self-blinding limitation honestly. The paused trial is not resumed by this framework revision.
Evidence And Remaining Validation
Component planning and purpose-led selection: 
Still to do: verify exact candidate identities against bottles; apply the documented scoring anchors and minimum quality gates to the recipe-specific brief; finish the relevant library holds; test role assignments and substitutions. No candidate here has been promoted to a proven ingredient.
Revision Record
September 12: retained the original candidate pool, separated Elijah Craig expressions, replaced deterministic grain/texture claims, added optional brightness, separated non-whiskey modifiers, and added evidence-status and trial requirements. Version 0.3 links the shared scoring and advancement rules without duplicating the rubric. No new pages, bottles or trial results created; the existing trial remains paused.