My 2 cents is that instead of preferring one blind shot over another, people should attempt to ID subsequent blind shots based on previous ones. This would show that the difference is systematic enough to be learned, whatever the preference.

I think it is misleading to talk about which alternative is better until one is confident they are intelligibly different.

From what I understand of the test, it never got to that point; the coffee sucked so badly that focusing on the systematic taste differences was impossible.

RapidCoffee wrote:If Jon & Co perform a blind taste test using paired Roburs and a GS/3, I'm willing to accept their results over anything I could currently do in my kitchen. Or likewise, anything you could do in yours, unless you also obtain paired grinders, machines that can consistently replicate brew temp and pressure, and set up a blind tasting protocol.

Hi John,

my findings are obviously good enough for me for my own personal choices, but I honestly couldn't agree more: in terms of value for the HB community as whole, the results of thorough blind testing by one or more groups of people, using contrasting beans and extractions suitable for the beans in question, experimenting with different espresso styles etc. would obviously be of much greater value.

Oh... and your laser diffractometer is in the post

Cheers

Mike

another_jim wrote:My 2 cents is that instead of preferring one blind shot over another, people should attempt to ID subsequent blind shots based on previous ones. This would show that the difference is systematic enough to be learned, whatever the preference.

I think it is misleading to talk about which alternative is better until one is confident they are intelligibly different.

From what I understand of the test, it never got to that point; the coffee sucked so badly that focusing on the systematic taste differences was impossible.

Very good point, Jim. I can understand John questioning my ability to consistently distinguish the SJ from the Elektra, just as I can understand Jon wanting to try blind tests of his own to see which he prefers, or you questioning the repeatability obtainable with a lever machine in such comparisons. Nevertheless, the one blind test I was able to do was between the two grinders, and much to my surprise (honestly, John!) it was always clear which grinder was which, whether they were single-dosed or not. I was always running two singles for this purpose, curious apart from anything else about how the flat/conical differences worked out - and to test my perceptions.

I think being able to distinguish consistently would be more appropriate than which is "preferred" (though it might be worth tracking that too!), and once I've roped enough victims in here after the Nino has arrived that is what I shall try to do, assuming people don't think the lever machines are totally invalid for the purpose. I find them both consistent if one is careful to manage them (the Cremina more than the Elekra, however - but I acknowledge that the bigger the group, the more difficult this will be to achieve.

Maybe we could use the Caravels...?

Cheers

Mike

another_jim wrote:My 2 cents is that instead of preferring one blind shot over another, people should attempt to ID subsequent blind shots based on previous ones. This would show that the difference is systematic enough to be learned, whatever the preference.

I think it is misleading to talk about which alternative is better until one is confident they are intelligibly different.

No question about it, however the fact that a particular test may not be capable of measuring a difference does NOT mean that a difference does not exist.

+1

RapidCoffee wrote:If Jon & Co perform a blind taste test using paired Roburs and a GS/3, I'm willing to accept their results over anything I could currently do in my kitchen. Or likewise, anything you could do in yours, unless you also obtain paired grinders, machines that can consistently replicate brew temp and pressure, and set up a blind tasting protocol.

+1
Paired grinders should be a must. IMO, and I defer to those that have used both, an E and standard maybe too different as evidenced by the 11 to 1 preference.

drdna wrote:No question about it, however the fact that a particular test may not be capable of measuring a difference does NOT mean that a difference does not exist.

But it could* indicate that the neither method produces a distinguishable difference. In which case either method is viable.


*I used could. I did it on purpose. It's cause I meant could.

cafeIKE wrote:Am I missing something

This is the reason I was reluctant to post the results. I only did finally post the data to illustrate that I had not allowed any preconception or my emotion to play any part in the test, as evidenced by my tossing out the whole thing as invalid. Other than that, the numbers are meaningless and any further interpretation would be out of context.

cafeIKE wrote:Paired grinders should be a must. IMO, and I defer to those that have used both, an E and standard maybe too different as evidenced by the 11 to 1 preference.

To assert that the grinders were too dissimilar is, well, ludicrous (IMO). Same age, motor, burrs, grind path...and we used WDT for both to remove any differences in distribution caused by doser vs. doserless.

If you think the grinders need to be even more identical then you will just have to purchase a pair yourself and break them in together. To carry it to the extreme, you could just state that the manufacturing+assembly variance on consecutive serial numbers is too much for this test and so there is no way to actually perform any comparison test for loaded hopper vs. single shot dosing.

But now we're a LOOOOOONG ways off from the "no brainer" implication, aren't we?

The shots tasted bad. Plain and simple.
The tasters had difficulty distinguishing any differences in the cup.
The judging was made based on least offensive...but all the test shots were sinkers.
I feel badly for the tasters, it was a painful task

The test "results" are nothing more than a coincidence, and if I had properly relayed the whole story or if you had been here you would probably agree (or maybe not)

cafeIKE wrote:IMO, and I defer to those that have used both, an E and standard maybe too different as evidenced by the 11 to 1 preference.

[EDIT: Wrote this while Jon was writing his post, so I didn't see it, but FWIW:] I don't think so. Motors and burrs are the same; the Robur E does have a propeller over the burrs that feeds (and likely pre-smashes) the coffee, which could play into things. I thought about having it taken it off, but I didn't think we had a short enough bolt to run without it, so I didn't even mention it to the other guys. I'm skeptical that it makes a difference, but we can look into taking it off next time, maybe.

I think the 11-1 preference is a heavy overstatement of our results. The shots were very close in characteristics, and we basically ended up rating them on the least presence of the defect of unpalatable sourness. In the room, the casual conversation was surprise at how very close the grinders were performing and the shots were tasting. I won't feel happy about gleaning much other than skepticism (and incentive to look further into this, as a long-time opponent of single dosing) from our tests, so they will have to be redone.

For what it's worth, Mike, I was impressed to see the Robur do so well at single dosing, purely from a visual/consistency basis watching him dial it in and then check it. Jon WDTs, which may explain his comparative success (compared to mine when I had that grinder). I'd tried a little bit of single dosing on the Nino some time ago with a very negative usability experience--bad static that makes the grinder look like cookie monster's lap. You have to rap the chute to get everything out, in my experience, and it's just a pain. If I single dosed, I would go back to a dosered grinder, that's for sure. Jon inspired me to look into it again last night, but after playing around again for 45 minutes or so with the Nino slingle-dosing, I realized I was very happy the way I already use it, very unhappy with the usability issues that go with single-dosing, and had no way to meaningfully compare the results of doing either for that specific grinder; I decided we'd wasted enough coffee for a day, and I don't think I'll be revisiting single-dosing on the Nino.

Just a few thoughts... I'm [secretly not] sorry I missed the fray on the testing that went on today. Good evening, gentlemen.

JonR10 wrote:To assert that the grinders were too dissimilar is, well, ludicrous (IMO). Same age, motor, burrs, grind path...and we used WDT for both to remove any differences in distribution caused by doser vs. doserless.

If you think the grinders need to be even more identical then you will just have to purchase a pair yourself and break them in together.

Halfway through the testing, switch grinders: take the one that you're using for single dosing and fill the hopper, and do the opposite with the other grinder.


-Andy

RapidCoffee wrote:Your assumption that the particle size distribution is the same is unwarranted speculation. My unwarranted speculation is exactly the opposite: changing the distance between the grinding burrs will almost certainly change the particle size distribution. The fact that the bean load is changing by a factor of infinity in single shot dosing (from ~16g to 0g), but only by a small factor with a bean load, makes this a near certainty. Only particle sizing (e.g., laser diffractometry) will tell us for sure.

Well, then help me understand how it's possible to dial in shots of identical dose with identical extraction characteristics (volume, brew ratio and time to blonding), and have different particle size distribution? If the particle size distribution is different, then what other factor could possibly make the extraction characteristics the same?

Of course, my assumption that the grinders were dialed in to make the extraction characteristics identical may be wrong. For example, if the cups had different brew ratios, I suppose the particle distribution could be different. But if that's the case, then I don't understand how the experiment could be of any value. Seems to me that all extraction parameters must be identical to determine if there's any difference between single dosing and loaded hopper dosing.