Nikka Pure Malt 17yo
Taketsuru Whisky | 43% ABV
Blended malts both cheap and dear
Continuing a recent theme of ridiculously expensive and otherwise difficult to obtain whisky, let’s look at an old Japanese malt blend and also ways we might get close to the same flavour experience without remortgaging any properties.
How do we plan on getting the same flavour profile you may ask? Well, by doing the same thing as our friends at Nikka did; blending. I’m absolutely, unequivocally and axiomatically not a blender, but having had the opportunity to speak to a few people that are, as well as work in or alongside a few micro-distilleries here in South Australia, I can say one thing with absolute confidence - it’s a seriously tough job.
It’s one thing to have a few open bottles and muck around mixing different ratios of things together, but not exactly outcome oriented. It’s an entirely different thing to have a specific flavour profile handed to you and then, with whichever blending tools exist at your disposal, combine them in a mind-bending superposition of art and science to bring forth that goal from the ether.
The number of variables that must be balanced at the same time while not losing sight of the larger picture of the overall flavour profile is quite daunting. Let’s ponder a few of those intricacies before utterly brutalising an attempt ourselves.
1. Goals
One ought to deeply understand the specific goal of creating a blend. A common example is a core range that a group or distillery has been making for a long time, say for instance something as well known as Glenfiddich 12yo. In this instance, a number of parameters have already been set for the blender; the minimum age of the stock they’re working with, the distillery profile, the final ABV etc. The blender will also be working from a large backlog of bottlings as reference points; bottlings from different eras which demonstrate which ranges of organoleptic properties the final blend should fall within. In this case there will also be a baseline recipe the blender ought to work from too; how many first fill versus second and perhaps third or even fourth fill casks, STR/decharred-recharred, how many should be bourbon versus sherry or some other previous fortified type. This will set a starting point which, especially when averaged over large batch sizes, should get the blender to within a very narrow margin of the desired outcome.
Another example is a blended whisky which needs to pull resources from a much larger pool; say for example Johnnie Walker Black Label. Again, in this instance there will still be many set parameters, previous examples to refer to as standards and a recipe guideline. The trickiness in this setting though is the variability that comes from not completely controlling all of the input variables; say a contract with another company is ended and their distillery components are no longer available for use. Even internal supply changes can affect resources; say for instance (all of this being completely hypothetical) Diageo HQ decided that Caol Ila is on a strong upward trajectory as a single malt. Of course they could restrict the amount of liquid which leaves the Diageo supply chain, but if the internal demand is great enough then it might be decided that only a fraction of its previous use in blends will be allowed going forward.
In this instance, it’s up to the blender to adjust the blend recipe using other resources to make up for the shortfall. Of course, by substituting a different peated malt to make up the phenolic profile of the blend, other variable parameters will also be affected, and so more changes may be required to adjust those back to specified requirements.
2. Alchemy
Next, there are physicochemical components to whiskies which, when blended together, will equilibrate and form the final structure and chemistry of the blend. Many of these are measurable and probably worthy of serious consideration, having a direct impact on how the consumer perceives the whisky organoleptically.
I’ve discussed many aspects of these in previous articles, but it’s worth reviewing them briefly. Primarily, there are fairly basic measures like total sugars, which should be obvious in their impact, and pH/titratable acidity. The latter, though, not only affects the sensation of tartness on the palate but also helps to determine chemical structures -for instance, acetic acid’s role in forming and stabilising micelles - and helping to determine the volatility of certain compounds such as pyridines and other basic molecules which can become bound in the liquid phase as protonated cations in sufficiently acidic solutions.
Other factors which might be monitored for consistency include tannins, both hydrolysable and non-hydrolysable (particularly if wine/fortified casks are heavily utilised), organic acids, alcohols, esters, phenols and other volatile components which dictate significant characteristics of the whisky. These factors would presumably be more closely monitored by large companies/producers with labs, or those with the budget/connections to send samples for external lab analysis.
3. Creation
In the case that a blender is responsible for creating a new product, they must either formulate a theoretical profile which they wish to achieve or create a number of different profiles to choose from in a more trial-and-error style formulation. In either instance, the blender is also responsible for ensuring that there are adequate streams of the desired components to be able to make that blend consistently going forward.
In certain instances maintaining consistency can be made easier by relying on wood technology to occupy a larger fraction of the blend’s flavour, minimising perceivable variation due to distillate components. In any case, there is usually a fair degree of freedom in the creative design, but also the crushing pressure of having that product represent your professional abilities and preferences.
Review
Nikka Pure Malt (Taketsuru) 17yo, 43% ABV
$600 AUD minimum
A big shoutout to my friend Jason for loaning me this bottle to be reviewed; there’s absolutely zero chance that I would have had the opportunity to try this otherwise. Even if a bottle was available from a local vendor, it’s just not the kind of whisky I’d go out of my way to pick up at the given price. Not even for my whisky club.
That said, I do have a history of enjoying the Nikka whiskies, particularly their Yoichi and Miyagikyo. I vaguely recall having tried the NAS and 12YO Taketsuru at various points too and feeling they were a pretty good representation of their single malt components.
Let’s see how they went blending some older stock, apparently with a significant sherry cask proportion in the mix.
Nose
Starts a little sulfurous and Caribbean rum-like; fermenting fruits, mild rubber, touch of solvent-plus-pineapple, pickled onion and a little methanethiol (think natural gas fumes). Presumably this has seen significant sherry cask maturation, though Yoichi is prone to organosulfurs too. Moving past that there are more stinky tropical fruits geared particularly to guava, passionfruit and grapefruit juice, then the more typical sherry cask stewed plums, dates and raisins. The casks continue with dried orange, roasted coffee beans and tobacco. There’s also a good maltiness, mild vanilla custard, baking spices and rather subtle smoke like a jacket worn to a campfire several weeks previous.
Palate
The sulphur is better behaved, orienting more strongly to the same tropical fruits, particularly passion fruit and grapefruit juices. More firm maltiness plus a sherry profile which hasn’t been overdone; fairly mild rancio, only a moderate sweetness, more pyrazine notes a la coffee and tobacco, orange peel, balanced cask spices and just a soft back palate bitterness with pleasingly light tannins. The distillate sings here as well with a bit more of the characteristic Yoichi fermentation profile; more juicy fruits and just a hint of pleasant yeastiness in support. The peat is integrated well here too, summoning more wood smoke, smoked salmon and lightly grilled meats.
The Dregs
I started off a bit unsure on first pouring this, particularly because of the sherry casks’ influence, but with time and a little of the sulphur blowing off this revealed a well-matured blend of characterful distillates which I really enjoyed.
I think the experience was slightly better blended for the palate than the nose, but that’s probably just personal preference given my aversion to too much sulphur. How does it fare overall? Well, based purely on organoleptics I think this warrants at least a good 7, realistically closer to an 8 if we allow for most people having a higher sherry affinity and a lower sulphur sensitivity than myself. But the pricing on this bottle, especially given the presentation, is unacceptable. At release in 2010 it was probably an excellent proposition and Japanese whisky does seem to be settling down and coming back to a semi-acceptable point, especially for younger stocks.
Further, I still think the Yoichi NAS in particular is relatively good value considering the calibre of liquid in the bottle. But these old bottlings with age statements in the current market? Forget it; this is still firmly geared to the luxury market and not meant for us mere mortal drinkers. For that it loses a point, as reflected below.
Score: 7/10 TK
Review
Various, 43% ABV
Average bottle price ~ $125 AUD ($110 AUD paid)
Right, so now that we know the profile we’re shooting for, what gets us into the same ballpark as that highly individual Taketsuru? From my view the most prominent aspects to target are;
Sulphur components, especially medium-high thiols oriented to tropical fruits, alliums and meats.
Prominent but balanced sherry casks, not too sweet on the palate, some age character demonstrated.
Medium tropical esters.
Good malt with non-medicinal peat featured but not pulling away attention.
A well-balanced mouthfeel which demonstrates sherry but isn’t too heavy, medium-low alcohol feel and intensity, not too drying and not too effervescent. Designed to make certain drinkers say “Smooth”.
Trying to achieve all of these while feeling integrated will certainly be challenging, especially without using any components which are too heavily priced; ideally I’d like to keep average bottle value between $100-$150 AUD if possible, given that seems to be a pretty good maximisation for the graph of diminishing returns domestically.
The bottles I’ve selected for this each feature at least two of the characteristics described in the above list (I’m guessing most of you can evaluate which is which) and the ratios obviously reflect how much of each feature I felt should be represented in the final blend.
Starting off, I used 5ml from each bottle to set a baseline for the overall profile. It was globally in a similar-ish place to the Taketsuru, but the equaliser was off; the sherry profile was pretty good so far as intensity, but a little too much of the primary fruitiness and not enough tertiary pyrazine complexity. There also wasn’t enough sulphur, particularly the alliums and grapefruit/passionfruit, and the tropical fruits needed more lift. The peat level was also pretty good but could use a touch more lift on the dirtier end of the spectrum. There also wasn’t quite enough malt coming through; again, the sherry fruits were a touch too prominent.
Accordingly, in the first round of adjustments I added another 5ml of Clynelish for the malt and fruity ester components and 10ml of the Secret Orkney for the tertiary sherry, thiol and peat components. This brought the whole much closer to where I wanted it, but it still lacked some touches of malt, tropicality and the sherry was still just a little too prominent. To adjust those three parameters, my last addition was another 5ml of the Ben Nevis, and this seemed to do the trick. I calculated the strength based on what had been added and found that it should have come out to around 48% ABV, so I gradually added roughly 6ml of water to bring the concentration to about 43% ABV and a total volume of roughly 55ml. I then split the blend into two 30ml sample bottles and set them aside to marry for a period of several days so they could settle and come to equilibrium of components.
Lastly, to test my handiwork I asked my wife to blind pour the two sample bottles and a measure of the Taketsuru into three identical glasses without telling me which is which (ie the basis for a triangle test). I strongly doubted that my homemade blend would be indistinguishable from the Taketsuru under blind conditions, but it never hurts to reduce bias, and this way I could establish a measure for how close the blend recipe was to the Taketsuru as a function of my confidence in discriminating between samples. The results were fairly predictable; it was indeed quite easy to (correctly) pick out the Taketsuru, but the samples weren’t complete night and day. The noses shared more similarities comparatively, while the difference in palate composition was somewhat better defined.
Moreover I enjoyed the blends pretty well; it helps that the components were mostly of a pretty high standard, but I was still pretty pleased with the outcome. Let’s review.
Nose
Some nice yeasty sulphur (mostly distillate derived but some light cask type in the mix too) components and a pinch of juicy sherry a la crushed cherries and mixed berry jam. There’s a more pronounced sense of American oak and youth compared to the Taketsuru too; butter menthol, vanilla and more pronounced baking spices. The peat’s a little more pronounced too, though I believe that to be synergistic through differences in both intensity and individual profile/expression of the peat.
Palate
Definitely a bit younger with the maltiness standing out a bit more clearly. The sulphur component feels pretty close except for the comparative shortfall in fruitiness; the average age difference is significant. There’s also a bit less sherry influence in the mix, although it’s possible that part of that perception is the oak types used (ie the Taketsuru might employ more European oak based on what I’m tasting). The texture here is a bit oilier too, though whether that’s due to the dominant lack of chill-filtered components or the contribution of an individual distillery - ie the Clynelish - is unknown, though I’d guess both.
The Dregs
Ultimately, this doesn’t do the same thing as the Taketsuru in the same way that the peated Glenturret didn’t replicate the Longrow 21 in one of my last reviews.
That being said, it’s perhaps within a standard deviation for flavour so to speak, but more easily more than a standard deviation cheaper by price. Had I used one or more components which better replicate some of the fruit+sulphur distillate notes of the Taketsuru, say for instance Springbank, Kilkerran or Yoichi - though that would really be cheating, and perhaps some older European oak sherry cask malts too, this could have probably landed much closer to the mark.
That would somewhat defeat the purpose of the exercise given the general lack of availability and/or pricing of those components in most areas. We should also remember that this was only a first attempt, with all the adjustments made in one day using only six component malts. Iterating this process using more bottles, a number of different starting points, more revisions and possibly even more palates (remembering that sensory panels by definition must utilise multiple nosers/tasters) could probably produce something much closer than my paltry attempt for about the same average bottle cost. Which brings me to these questions to finish on.
Which bottles have you tried to replicate? What did you use, and how close do you think you got? And to the real blenders out there that actually know what they’re doing; what tips would you offer to us amateurs?
Score: 6/10 TK
Share your thoughts in the comments below.