Tavern Ale, Version A
What version A is
This is a remembered recipe, not a recovered one. The original formula was never preserved — the historical record says so plainly. What's below is Craig's recollection, specific where he's sure and flagged where he isn't. Version A exists so there's something real to brew and argue with. Brew it, improve it, and tell us what you found — version B should come from someone else's kettle.
The story
It came out of frustration. Every batch Craig brewed during that stretch went to production testing, which meant he never got to drink any of it. So one day, mid-research, he slammed together a beer for himself. It turned out to be the best thing he ever brewed — good enough that it went to the beer company, and good enough that Bard's first investor wrote a check after tasting it.
Nobody planned it. That's most of why the formula was never properly written down.
What it was supposed to taste like
A light IPA carrying honey and a slight toasty, nutty aroma and flavor. Amber. High drinkability. The thing that made it work was the interaction: the toast, the honey, and the hops balanced each other — bitterness just under a light IPA, deliberately not enough to bury the honey.
Version A — the recipe
Tavern Ale, Version A
- Protocol
- GFB-R3
- Version
- Version A
- Effective
- August 2026
The best beer Craig ever made, reconstructed from memory: a light-IPA-strength amber ale carrying honey and a toasty, nutty depth from roasted buckwheat. The toast, the honey, and the hops balance each other — bitterness deliberately not enough to bury the honey.
Memory: This is a remembered recipe, not a recovered one — the original formula was never preserved. Specific where Craig is sure, flagged where he is not: the honey quantity is the guess (somewhere between 1 and 4 cups), and settling it is Research Ask #2. Version B should come from someone else's kettle.
Ingredient bill
| Ingredient | Amount | Basis |
|---|---|---|
| Sorghum malt, Vienna roast (house standard) | 9 lb — 90% of grist | Memory |
| Roasted buckwheat groats (sold as kasha) | 1 lb, medium-to-dark roast — 10% of grist Plain unroasted groats will not give you the same beer | Memory |
| Honey | ~2 lb (2½–3 cups), at flameout Estimated — the one number nobody recorded; reasoning on this page | Guidance |
| Hallertau Hersbrucker | To ~28–30 IBU: bittering plus a late aroma charge Variety remembered; schedule reconstructed | Memory |
| Wyeast British Ale | Pitched and held at the bottom of its range A couple of degrees cooler than standard — deliberate | Memory |
| Water + gluten-free yeast nutrient | Standard normalization for your supply, built for the yeast The original ran Kansas City water | Memory |
Stage before brew day
the Reliable Mash, with the buckwheat in the grist
- Do
- Run the Reliable Mash scaled to the 10 lb grist, kasha included.
- Add
- Per the Reliable Mash run sheet
- Advance when
- Wort collected.
- Record
- Gravity · Mash behavior with 10% kasha
Hersbrucker, two charges
- Do
- Bittering charge toward ~28–30 IBU total; late charge for aroma. Nutrient per supplier.
- Add
- Hersbrucker bittering charge
- Hersbrucker late aroma charge
- Gluten-free nutrient
- Advance when
- Knockout.
- Record
- Actual charges and timings — the schedule is reconstructed; yours becomes data
Flameout — protect the aroma
- Do
- Add the honey at flameout (or in the whirlpool, cooled below ~180°F). Record the exact amount.
- Add
- Honey, ~2 lb — record what you actually used
- Advance when
- Dissolved and in.
- Record
- Honey amount and timing — Research Ask #2 lives here
British ale, run cool
- Do
- Pitch and hold a couple of degrees cooler than standard for the strain. Let it finish fully — the honey adds fermentables.
- Add
- None
- Advance when
- Stable terminal gravity.
- Record
- Temperature held · Fermentation curve · Final gravity
Package and judge the balance
- Do
- Carbonate and taste against the target: light-IPA drinkability, honey in the nose and finish, toasty-nutty depth underneath.
- Add
- None
- Advance when
- In the glass.
- Record
- Where the balance landed — and what you would change
| Step | Action | Target | Add | Advance when |
|---|---|---|---|---|
| 1. Mash: more detail | Run the Reliable Mash scaled to the 10 lb grist, kasha included. | the Reliable Mash at Homebrew Scale Per the sheet |
| Wort collected. |
| 2. Boil: more detail | Bittering charge toward ~28–30 IBU total; late charge for aroma. Nutrient per supplier. | Boiling 60 min |
| Knockout. |
| 3. Honey: more detail | Add the honey at flameout (or in the whirlpool, cooled below ~180°F). Record the exact amount. | Flameout, or whirlpool below ~180°F — |
| Dissolved and in. |
| 4. Ferment: more detail | Pitch and hold a couple of degrees cooler than standard for the strain. Let it finish fully — the honey adds fermentables. | Bottom of the strain’s range To terminal | None | Stable terminal gravity. |
| 5. Finish: more detail | Carbonate and taste against the target: light-IPA drinkability, honey in the nose and finish, toasty-nutty depth underneath. | Amber, drinkable, balanced — | None | In the glass. |
Run each step with its checks and context
Use the run sheet during a routine mash. Use these expanded steps when setting up the process, training an operator, documenting a deviation, or diagnosing unexpected behavior.
the Reliable Mash, with the buckwheat in the grist
Objective
The remembered mash is the standard protocol; the kasha rides along in the grist for flavor, not conversion.
Operating instruction
Run the Reliable Mash scaled to the 10 lb grist, kasha included.
- Per the Reliable Mash run sheet
What good looks like
Wort collected.
Watch for
No additional step-specific warning.
Record
Gravity · Mash behavior with 10% kasha
Level 3 · Mechanism and evidence the Reliable Mash at Homebrew Scale
Hersbrucker, two charges
Objective
Bitterness just under a light IPA, with real late aroma.
Operating instruction
Bittering charge toward ~28–30 IBU total; late charge for aroma. Nutrient per supplier.
- Hersbrucker bittering charge
- Hersbrucker late aroma charge
- Gluten-free nutrient
What good looks like
Knockout.
Watch for
No additional step-specific warning.
Record
Actual charges and timings — the schedule is reconstructed; yours becomes data
Flameout — protect the aroma
Objective
Honey ferments nearly dry; the perceived sweetness is aroma, and aroma survives only if the honey goes in late.
Operating instruction
Add the honey at flameout (or in the whirlpool, cooled below ~180°F). Record the exact amount.
- Honey, ~2 lb — record what you actually used
What good looks like
Dissolved and in.
Watch for
Boiling the honey hard — the aroma you were buying dies in the kettle.
Record
Honey amount and timing — Research Ask #2 lives here
British ale, run cool
Objective
The cool bottom of the range was deliberate — restrained esters under the honey and toast.
Operating instruction
Pitch and hold a couple of degrees cooler than standard for the strain. Let it finish fully — the honey adds fermentables.
What good looks like
Stable terminal gravity.
Watch for
No additional step-specific warning.
Record
Temperature held · Fermentation curve · Final gravity
Package and judge the balance
Objective
The beer works when toast, honey, and hops hold each other up.
Operating instruction
Carbonate and taste against the target: light-IPA drinkability, honey in the nose and finish, toasty-nutty depth underneath.
What good looks like
In the glass.
Watch for
No additional step-specific warning.
Record
Where the balance landed — and what you would change
The honey problem
This is the one Craig didn't pay attention to at the time: enough honey went in to leave a slight residual sweetness, and the amount was never recorded. Somewhere between 1 and 4 cups.
Here's the reasoning behind the 2 lb (≈2½–3 cups) estimate, so you can argue with it: honey ferments nearly dry, so it does not leave sweetness by sitting in the beer — perceived sweetness comes mostly from aroma, which survives only if the honey goes in late. At 2 lb in five gallons, honey is roughly 15–18% of the fermentables: enough to read clearly in the nose and finish, not enough to thin the body or push the gravity into a different beer. Added at flameout (or in the whirlpool, cooled below ~180°F), the aroma survives; boiled hard, it doesn't.
This is the single best thing to test. Brew it at 1 lb, 2 lb, and 3 lb and tell us where the balance lands — that's a real experiment with a real answer, and it settles version B.
Sourcing the buckwheat
Roasted buckwheat groats are usually sold under a different name: kasha. Search for "kasha" or "roasted buckwheat groats" and pick a medium-to-dark roast — the darker end is what carries the nutty, toasty depth this beer runs on. Plain (unroasted) groats will not give you the same beer. More on how the grain behaves: Buckwheat.
Why German hops
Craig's field experience, stated as such: German hop varieties worked well with sorghum malt, and some others didn't — Saaz in particular clashed. That's an observation from commercial brewing, not a tested finding, and it's exactly the kind of thing a few homebrewers could settle in a season. If you test hop varieties against sorghum malt, we want the results.
What to change, and what to report
Four things worth varying, in order of how much they'd teach us:
- Honey load and timing — 1 vs 2 vs 3 lb, flameout vs whirlpool vs fermenter.
- Buckwheat percentage and roast depth — 10% is the memory; darker or lighter changes the whole balance.
- Bitterness level — "just under a light IPA" is a feel, not a number; find where the honey starts to disappear.
- Yeast temperature — the cool end of British Ale was deliberate; test whether it matters.
Report gravity, fermentation behavior, and above all what it tastes like. A version that beats this one gets published as version B with your name on it.
Related
- Tavern Ale — the historical record and what it will not claim
- the Reliable Mash at Homebrew Scale — the mash this uses
- Buckwheat
- Homebrew: The Research Bench
Source and Validation Notes
Version A is Craig Belser's recollection of a beer brewed once, without a preserved formula, and published as a starting point rather than a record. Grain percentages, hop variety, yeast, water treatment, and mash are remembered; the honey quantity and all hop timings are reconstructed, and the reasoning behind each reconstruction is stated above so it can be checked. Nothing here has been re-brewed or verified since. The hop-variety observation (German varieties suited sorghum malt; Saaz clashed) is commercial field experience, not a controlled result.