Samsonite,
I have seen the 1995 article, but it is a summary of the original article here: http://www.franklinbrew.org/brewinfo/brauweltfwh.html
The original article provides some interesting insights. While it unfortunately does not provide the original hop schedule in its entirety, it does provide experimental levels for the original brews as well as rough hops amounts and schedule.
For brew A, we see that the overall bitterness of the FWH brew (by EBC) is 6.35% higher than the non-FWH. Brew B has a whopping 28.1% increase in bitterness for the FWH case. In both cases, late addition hops were moved forward to FWH additions. For A, 34% of the hops was moved forward while for B approximately 53% of hops was moved forward. Both were low alpha hops (Tettnang or Saaz) - likely in the 4 to 4.5% alpha range.
Looking at case B, lets make some very rough assumptions:
- For simplicity, lets make it a 5 gal brew, 6 gal boil at roughly 1.045 OG (which can be backed out from the numbers given)
- Assume the hops all Tettnang at 4%, and that the original recipe had 50% of the hops boiled for 20 min, moved forward to be the FWH
- Again for simplicity lets use 1 oz boiled for 80 min, 1 oz added for 20 min as the baseline
Using the hop calculator (Tinseth), I get:
1) 25.6 IBU for the original recipe with half the hops at 20 min, half at 80 min boil (recall that FWH measured should add 28.1% to this number by the EBC measure)
2) 32.5 IBU for the original recipe with all of the hops (2 oz) boiled for the full 80 min (no FWH adjustment) -- 26.9% above the baseline case
3) 35.8 IBU if I use a FWH adjustment of 10% -- 39.8% above the baseline case
4) 21.6 IBU if I use a FWH adjustment of -33.5% as you suggest - or 15.6% BELOW the baseline case
5) Alternately if I add 110% of the 20 min contribution (per your rule of thumb) I get about 26.5 IBUs -- 3.5% above the baseline
Looking at the actual experimental results, it is interesting to note that case (2) above comes closest to the original case, indicating that leaving the FWH in for the whole boil does not significantly change its contribution to overall bitterness. This does make some sense - the alpha acids in the hops do not disappear when we FWH, and in fact assuming you leave the FWH in the wort (i.e. it is NOT a mash hop but a FWH) those alpha acids will be extracted during the boil.
I also ran the numbers for case A, which is a substantially lower increase in finished EBC alpha (about 6%) even though the wort isomorized is much higher in the FWH (13% higher). Again, the zero FWH addition comes closest in the IBU increase to the experimental results, though it is slightly high for this case. Your recommended setting of -33.5% FWH adjustment comes in low by some 28% from the measured EBC value.
Unfortunately, we only have two experimental data points with a very large variation (22% difference in EBC values and a little smaller if we look at the Wort values) between batches A and B.
However, overall I would say a FWH setting of around zero to -5% would hit the midpoint of these two data points. The BeerSmith default of 10% is probably on the high side, but a setting of -33.5% is way too low. The hops are boiled the full time, plus steeped for some time, so it does make sense that their contribution would be around the value of a full boil hops.
I would love to see some actual IBU data from some more FWH brews. I know I have made a large number of FWH only brews with great success, and they have not been unbalanced.
VR
Brad