Method and Style:
I recommend the pick-and-add method of weighoff. Key balancing weights are used in pairs, upweight on the bottom, pairing weight on top. Add a pair (downweight) to engage the key and note the speed of hammer travel. Pick the pairing weight off (leaving the upweight) and note the speed of hammer return. These two steps reveal the least weight that will engage the key and the most weight the key can lift and still return to rest. Choose downweights and upweights that work well with the action and match speeds as you go by trial-and-error placement of one or more key "leads" (I now use copper) on top of the key to be installed later.
Traditionally, the trade used downweight-only in weighing off. This method was fast but had the weakness of obscuring inconsistencies in action friction, which translated to inconsistencies in upweight (arguably more important than downweight for good repetition and playability).
In the ‘70’s and 80’s, when I was starting out, some technicians (including myself) felt they could do better than the downweight-only method of weighoff. I would choose a series of upweights, starting with the least that was viable in the bass (20g was my minimum) and rising to the most I could get away with in the high treble (generally 25g – other specs being optimized). With sufficient friction in the system, I still find this to be the useful upweight range (although my 20-25g might be someone else’s 21-26g or higher, as I use quite a positive speed of key/hammer response with no action bumping).
I would then select a series of downweights ranging in grams from mid-50’s to mid-40’s, and I would plot out an arbitrary pattern to switch in lighter downweights and heavier upweights as I proceeded from bass to treble. I applied first downweight then upweight and tried key weights on the keys until down and up speeds matched. This worked well (and was more precise than the traditional approach) but contributions from two people significantly improved both results and speed of execution.
David Stanwood’s Balanceweight System and New Touchweight Metrology nudged the entire trade to better understand and better apply principles of physics and geometry to their piano action work. In particular, I have found his concept of balanceweight (halfway between downweight and upweight) to be useful.
The second influence came from Bruce Stevens, with his stack-of-two-weights approach, shifting my process from manipulating two sets of gram weights and one or more key weights (a process of picking up and putting down, over and over, that really asked for a third hand) to just picking and adding the pairing weight with one hand and positioning the key weight with the other. Brilliant!
The Grandwork Weighoff Kit supports two common styles of weighoff with this method. The first works for pianos with a consistent downweight (often around 50g). In this style, the pairing weight decreases 1g for every 1g rise in upweight (20g upweight + 30g pairing weight = 50g downweight, 21g upweight + 29g pairing weight = 50g downweight, and so forth).
The other style provides more downweight in the bass and less in the treble, relating downweight to hammer weight. The ideal balanceweight for an action offers a user-friendly collaboration between each note’s readiness to play and its push-back to the player while being played. In this style, the pairing weight decreases by 2g for every 1g increase in upweight (34g pairing weight + 20g upweight = 54g downweight + 20g upweight ÷ 2 = 37g balanceweight, 32g pairing weight + 21g upweight = 53g downweight + 21g upweight ÷ 2 = 37g balanceweight, and so forth).
Working from bass to treble, one style produces the same downweight throughout, while the other offers incrementally lower downweights. But the key is that both styles feature incrementally rising upweights, a consistency lacking in the traditional weighoff. And for ease of execution, you don't have to plot switches in downweight and upweight, just use the pair that gives the easiest-to-read symptoms. They will automatically produce a consistent transition from bass to treble.
When I install new hammers, I add two other factors. One is smoothly descending strikeweights, bass to treble, accomplished through taper and tail thickness adjustments. The other is to match the highest bass hammer with the lowest tenor hammer in weight. This adds consistency of inertia to consistencies of balance and friction. Minimizing geometry and regulation irregularities refines this further (custom boring new hammers to match the arc of string heights from a crowned plate, for instance, uncompromises string-height-related regulation discrepancies).
We now have incremental choices of knuckle and whippen heel placement to facilitate action geometry tweaking. Wessell, Nickel & Gross has brought us composite shanks of consistent weight and stiffness and hard bushings that allow both less center friction and more stability. Whippen heals for sharps can be 2mm taller to mitigate unavoidable geometry/leverage differences.
Trial-and-error still proves the fastest route to finding the best balance. Place a suitably sized key weight on top of the key and find the spot where adding and picking the pairing weight matches speeds of hammer travel and return, respectively. If a key stands out with slower speeds, there is extra friction, faster speeds, less friction.
Caveat. The installation of a key weight won’t work if there is one in the way. Assess the overall add/subtract trend that your regulating will induce and pre-remove key weights as needed. Drill out holes with an oversized bit (9/16” works well) and plug with a wooden bung. New bungs and key weights in the blacked areas of sharps should be painted black when done. Strategic pre-removal is time-consuming, but it makes weighing off easier and more accurate.
Another notable approach to weighoff comes from Ed McMorrow. As part of his Light Hammer Tone Regulation system, Ed way reduces strikeweight and removes all key weights from the keys, way reducing inertia in the system. The resulting 70g plus low bass downweight reportedly feels responsive rather than heavy and the lack of key weights really helps the player in fortissimo playing. He applies this exclusively to Steinway pianos but suggests a version of this method is worth considering for any grand piano.