shadowfax wrote:There are tons of home roasters out there all over the internet; what basis would you be have for assuming that you'd taken a good 'cross-section' of homeroasters?
I don't think you'd ever try to promise you'd take a truly representative sampling - just as you wouldn't try to do that for "US professional roasters." You'd simply have to live with the fuzziness on both sides.
That being said.... the approach detailed above (top home roasters from each round go to a "finale") would yield a good "State of Home Roasting" sample IMHO.
Of course, this doesn't address the "compare it to pro roasters" issue - but as I said earlier that would be an area I would personally shy away from.
shadowfax wrote:How would you link that to experience, and without a truly massive sample size, how do you identify a poor roaster from a guy who had a possibly one-off screw-up (maybe not so much the "one" part

)?
I don't think you would worry about experience.
In terms of sample size... I think 30 home roasters would be sufficient.
Each would provide 1 coffee, 10 would be cupped per round, top 2 advance.
Each of the top 2 would be asked to provide 2 coffees for the finale.
I'd say that would be worthwhile if not scientifically or statistically complete.
If someone sends in a screwed up roast - then I think they get what they deserve.
If you can't tell that you screwed up your roast, I'd have a hard time arguing that you know what you're doing.
shadowfax wrote:If you want to settle this debate, I think you'd have to be able to ID inconsistency vs. consistently poor quality, and reading the accounts from
Rate My Roast would suggest that orchestrating a bunch of home roasters delivering pairs of roasts that they're truly happy with, all for one big comparative taste test, would be a monumental task that would put asterisks all over the place on the test and make your conclusions either super-limited (not very useful) or controversial (not widely accepted).
See above model.
Seems like it would be valuable against what you're describing.
First 30 roasters sign up and commit to sending in coffee. If they fail to send it in - they get called out for it (incentive).
Top 6 end up providing 3 different coffees over 2 roasts (consistency).
The model of the "set of serious palates" for evaluation works (quality).
Results would likely not be "super-limited" (but would be somewhat limited).
Absolutely I can guarantee it would be "controversial" -- but it would still be TRUE (ie see the "controversy" of the SF results as it relates to Equator - Roaster of the Year; and Blue Bottle - cult favorite).
shadowfax wrote:I don't know if this debate will ever really be settled; again, I'm really skeptical that a huge test like you describe (which would be interesting to participate in AND to read about) would get us too much further to that end.
If you are talking about the home vs pro debate - as noted above this would not be designed to address that and it's something that I don't think is worth debating. I think everyone in the center of the bell curve agrees on that debate (a somewhat skilled home roaster is very likely to be able to sometimes produce coffee better than the "supermarket" roasters produce; a skilled home roaster is very likely to be able to consistently produce coffee better than these "supermarket" roasters and is likely to be able to periodically produce coffee better than all but the top 1 or 2 percent of professional roasters; very few home roasters are likely to be able to produce that is consistently on par with what those top pro roasters produce).
If you're talking about "how good is home roasted coffee" as the debate - I think something like this would give a sample in time.
And to be honest - I'd even be willing to predict how it would come out. I think that anyone who has been to any large home-brewing events / meet ups / conventions can probably apply the same % breakdown to coffee as you see in home-brew.