Hacker News Reader

The reason I recommended buying coffee within days of its roast is that you'll probably drink that same batch over the next week or more. Things start going downhill after around a couple of weeks off roast. So, if your coffee is 1+ week old when you buy it, it's probably going to be missing something for at least some of the days you drink it.

I drink mostly decaf these days, and it stales even faster than regular. The difference is so stark, that to get a consistent espresso extraction I have to adjust the grind setting finer as the coffee ages. If a 10 on the Niche Zero is where I dial in coffee 4 days off roast, I'll typically need to dial down to an 8 by the end of two weeks.

Still, I don't have any trouble believing the results of your experiment, or the YT video's experiment. Two weeks is still pretty fresh. But if you repeat your experiment with 4-day coffee vs. 4-week coffee, maybe your tasters will be able to detect a difference. (Also, depending on how you prepare it, 1-day coffee might be "too fresh"; it's still outgassing a lot. I'd suggest you try 4-day coffee as your control.) If you try decaf, I'd expect the difference to be more detectable.

Anyway, thanks for the empirical evidence.

Any experiment is going to leave questions, but unfortunately, not many people or places have the luxury of getting 17 people together many, many times.

That said:

> Things start going downhill after around a couple of weeks off roast

"a couple of weeks off roast" is exactly the test we did. It's reasonable to propose other ages and you could well be right that results would be different.

I think 2 weeks off roast is close enough to 'fresh' that the difference is going to be subtle. I'd love to see an experiment like this with 3-day old vs 2-months old. I'd guess most people could tell in that case.

And maybe then bring that two months in and see where the cutoff is that people can/can't tell. I'll guess it would be 3-4 weeks or so for most subjects.

I've never heard of a "triangle test" before but I like that format!

Yes. 1/3 chance of being right randomly. So even 3 right answers is 1/27, which is below the p = 0.05 threshold commonly used.

You probably don't want to test 1 day vs 2 weeks. At day 1 the beans have not degassed enough to release all their flavor unless you take special care during the brewing process. Around 1 week after roast is optimal for a regular v60 brew. 2 weeks isn't long enough for the beans to have gone (very) stale. After 3 weeks is when the taste really drops off.

What brewing method did you use, regular cupping procedure?

It was done in one of the Google coffee bars. I don't remember exactly how that works, but you can be sure it was done with professionalism

So you’re saying there should be a difference between 1 day and 2 weeks, but the google story said otherwise, basically no one could tell.

More accurately, not enough people could tell to be statistically significant. This is all explained on the video.

People who say "use 4 weeks" are right. It's just hard to organize a whole lot of these things. We're lucky we could do it once.

Most definitely, but it depends on a lot of factors, like how and when the beans and coffee were prepared. Overly fine ground (maybe with a blade?) dark roast in a filter pot? It's going to be terrible coffee regardless.

A high quality bean lightly roasted and ground perfectly just now in a pour over? Two days ago versus two weeks will be noticeably different, three to four weeks after it'll be obvious.

Edit: I'd like to know more of the variables that went into this. "Most people" I know don't drink black coffee, if they drink coffee regularly, or don't drink coffee at all. If you're used to drinking store bought grinds and using a filter pot (or some kind of pod system) then you're probably already a lost cause here.

You're right about black coffee (I use a teensy bit of milk & sugar myself, way less than a tsp), but for a test you want consistency.

Actually I think the last few days recommendation is off by quite a bit. Most boutique roasters I know about typically recommend to use the beans within 2 month after roasting and some even recommend to not use the before 1 or two weeks after roasting .

A different test would be the grinding one. I am very sure I can taste the difference between fresh ground and ground a day or two ago coffee (for espresso however). I would be interested what you get for that experiment.

Not an expert by any means, but given personal experience: By taste it feels fairly obvious when the roast is over 3-4 weeks old. And it's not only perceptible by taste, it is fairly difficult to ignore while blooming a pourover when it's fresh but specially so during calibration for espresso: when a grind is on its first week it's much harder to find the right spot, and when it's over ~3 you're clearly grinding noticeably finer.

Caveat: If you're unfamiliar with what "normal fresh" is for a specific bean of course it's going to be hard to tell at which stage it is but it doesn't mean that you wouldn't be able to state a clear preference if presented with the same bean before, during, and after its ideal degassing window.

And double caveat: A bean is not a bean is not a bean. Dark roasts degas much faster than light roasts and there's some wild variation out there in what's grown.

Spitballing from the basic chemistry, freshness is mostly about the time before the beans are rancid, the packaging used will hold in volatiles but oxygen + heat will make beans unpleasant. I wince a bit when I see coffee in the coop-style dispensers, that's a recipe for unsavory flavors.

So date-since-roast is also a proxy for likelihood of rough treatment, and it's lore from a world which includes roasted beans which aren't stored sealed in any way. I have had coffee taste fine a month after roast if stored carefully, and in the tropics, fresh roasted beans can go off on you in a few days, if they live in a kraft paper bag at outdoor temperatures and humidity.

You should try fresh, 30 days, and 60 days. The cafe I worked only adjusted grinds for coffee which was 30 days after the roast date. Allegedly this was decided through many many taste tests.