The matrix

[flash_uk]

I have embarked on a little mission. Something tells me others may have journeyed down the same path...so I thought I'd canvas comment, advice and opinion

As a recent convert to the pleasures of port, I find myself with a cellar sparsely populated with port, and many fun years ahead changing that. I am on a huge learning curve about which years and which shippers would be worthy additions to the cellar, and to help build this knowledge I have created a spreadsheet grid, with years and shippers, which I am populating to give me an idea of what is worth buying.

To populate the matrix, I am cycling through years of tasting notes and forming a view on each year/shipper. I am slowly adapting and refining exactly what info I record, and so far I have concluded this would be helpful:

Rating: G, G/VG, VG, VG/EXC, EXC ...so a 5 point range. Also considering tracking consensus range of scores e.g. 89-92, but obviously not always possible to do this

Status: Keep, drinking nicely now, drink up, fading, indeterminate, avoid (incl reason e.g. TCA problem)

Next review: year in which should next try the juice (so if it is young and needs 5 years before next taste, this is 2019)

Thoughts anyone? Other info that would be useful to track?

[flash_uk]

A quick addendum...I have read with interest the many lively threads on scoring systems, merit of scoring etc, a few here:
Scoring Ports
Uncle Tom's Scoring System
The best intentions and futility of scoring port

I'm not really inviting a rerun of all that...just asking, if you were recording some views on a year and a shipper, what would you include? Whatever I end up doing is for my view, so if I pick a scoring system, as long as I understand it and am happy with it then great!

[griff]

I have come to the conclusion that, ultimately, communion with port is personal. One man's meat is another man's poison and all that. You can crowdsource opinion on ports that many are able to enjoy, but what one finds beautiful is personal i.e. Plato was wrong

WIth that out of the way I think one can better rate ports on their attributes rather than their quality. So you could rate port on it's readiness to drink (a reason I like Tom's scale in preference to the 100 point scale), as well as acording to their character. I know I like more savoury ports so I lean towards Dow and Warre ports. Spicy styles like the Portuguese houses are then sought, followed by fruity, powerful ports like Taylors. Of course, individual vintages and individual bottles refuse to be generalised in such a manner. However I shall not let such occurrences dissuade me from categorising port house styles, so I shall continue to expose myself to vintage port for scientific purposes of course!

[jdaw1]

Flash: it isn’t really obvious what you want, other than everything. Do you want to be told that ’63 is showing its age more than ’55 or ’66 or ’70? Do you want to be told that Cockburn was great from 1847 to the 1960s, then fell apart, and is now in good hands? What do you want?

[flash_uk]

it isn’t really obvious what you want

Guilty m' Lord

Agree with griff that port is v personal and I have no desire or energy to try and reduce the merits of wine to a number being an average of averages.

As to what I want, really just something that for now can give me a rough feel for a year/shipper. As an example, on my pretty coloured matrix I have RP80 marked as "avoid", and Ni94 as "systemic problem - avoid", GC83 as "maybe past best". For the 94s I marked F94 as Excellent and GC94 as Good, on that 5 point scale mentioned above. maybe others can remember this stuff easily in the head, but I can't. So when I look at something on an auction site e.g. a C75 estimated to go for £275/c12, right now I am at square 1 as to whether that might be decent value, likely to taste awful, WOTC for Cockburn etc etc...

Ultimately, all of the opinions on what the juice was like when tasted and what prospects it might have, would be my own opinions, but in the meantime while I have yet to taste many, if I can get half a sense from others notes, then fantastic.

Given all of that, I have been tending to the view that, tracking some kind of view on whether the juice was good and what it's prospects might be (akin to Tom's method), a simple view on where it is in the cycle (drink, keep, accepting that there is not necessarily a universally correct linear sequence for this...only an opinion again), and then the "when next" date thing to give me a nice way of picking something to drink if I am at a loose end.

Would be interesting to hear if others have previously attempted something similar and found it useful, or have never and will never try (DRT?) or abandoned it - maybe found it didn't really help in the long run, or was too simplistic, or too complicated.

[jdaw1]

The best 1975s (Croft, Noval) are not great. Most are rubbish. Cockburn ∈ Most.

(Specific questions work well.)

[PhilW]

My starting point is to give an overall rating to each bottle I taste: excellent, very good, good, ok, poor, faulty
Since I want to see some relative comparison and at some tastings all bottles can be in the top three categories, I allow myself use of + and - as a modifier to each as well. I also break my own rules occasionally and allow an outstanding, or a dreadful. Potentially, this above could map onto a points score (or some ranges) but certainly doesn't provide the detailed discrimination that a 100pt scale using half-point resolution can, but I'm not sure I could provide reliable absolutes at that granularity, so this method suits me well.

Sometimes I find it difficult to define a common absolute range across different categories, so will suffix the rating, for example "good (for LBV)"; this generally happens in the lower part of what would otherwise be an absolute scale, where differentiating between several bottles I want to label all as "ok" or similar.

In addition to the overall rating, I add other notes as needed; I often also note a value for money rating (using the typical bottle cost for purchase, not how much I paid), as well as level of maturity. Equally I might note that it is very good but not quite my favoured taste. So overall, for each bottle I taste I record:

Date/location,
Bottle (category [VP, tawny, colheita, white, LBV], vintage, house, size)
Bottle source (including cost and when purchased, if one of mine)
Rating (excellent, very good, good, ok, poor, faulty) +/-
Rating comment (maturity, VFM)
Tasting notes / General comment (Sometimes colour, nose, taste, or any comments)

Regarding the tasting notes, I know I can sometimes be poor at naming flavour elements, so I tend only to post tasting notes for ports not shared with others who can do that far better than I, and for bottles which I think others might be interested in (e.g. you probably don't need yet another tasting note for W77 whenever I open a bottle but hopefully the recent CkQdC92 note might have been interesting as it less common.

In the end, I use my notes to help remind me about details of previous events, as well as to guide future purchasing and drinking. It would probably be useful to have these in a database form, but I use a simple excel spreadsheet which allows me to use the simple sort mechanism to easily list any house and/or year combination for all bottles I have tasted, when needed; that works well enough for my needs.

[djewesbury]

In reply to JDAW, I think Flash is looking for tips / insights on general scoring and notation methods that we have found helpful / useful, rather than specific comments on shippers / vintages.

My own comment in addition to what has already been said is that there is far too much bottle variation to reply say that X shipper's 19XX is always worth buying / avoiding, with a few obvious exceptions (for my money, G70, whenever and wherever).

[DRT]

I think I have found exactly what is required.

[djewesbury]

I think I have found exactly what is required.

I seem to remember suggesting this a while back. But people do like systems.

[flash_uk]

I think I have found exactly what is required.

Ha ha! yes, yes...can't deny these are useful. Need to read these each time you want to know about one. My desire is to get a simple view on one page!

[DRT]

My desire is to get a simple view on one page!

Landscape = 50+ Shippers x 200+ Vintages, Portrait = 200+ Vintages x 50+ Shippers. A0 should do nicely

[flash_uk]

My desire is to get a simple view on one page!

Landscape = 50+ Shippers x 200+ Vintages, Portrait = 200+ Vintages x 50+ Shippers. A0 should do nicely

I'm currently on A3 with 27 shippers and 45 years. That is covering back to 1945 for a good chunk of shippers. For those not on the page I'll use your suggested solution on an as-needed basis.

[flash_uk]

My starting point is to give an overall rating to each bottle I taste: excellent, very good, good, ok, poor, faulty
Since I want to see some relative comparison and at some tastings all bottles can be in the top three categories, I allow myself use of + and - as a modifier to each as well.

I like the idea of a + or - modifier!

[PhilW]

My starting point is to give an overall rating to each bottle I taste: excellent, very good, good, ok, poor, faulty
Since I want to see some relative comparison and at some tastings all bottles can be in the top three categories, I allow myself use of + and - as a modifier to each as well.

I like the idea of a + or - modifier!

In reality I really only use F, P, O, O+, G, G+, V, V+, E, E+ and E++/+++ exceptionally ([F]aulty, [P]oor, [O]k, [G]ood, [V]eryGood, [E]xcellent).

I did consider a similar matrix idea a while ago; after much musing, the best I could think of to allow enough information yet keep it simple to present was:
- Each horizontal line is a house
- Each vertical is a year
- Each cell provides space for up to four blobs (simpler would be one blob) arranged in a square
- Each blob is a graphic, perhaps a circle (for four blobs, or glass if only one blob), which can be filled any amount from empty to totally filled
- Each blob would represent a tasting of that wine/house (so if only tasted once, there would only be one blob; if tasted >4 times, they would represent the last four tasted)
- Each blob is filled based on the rating from the tasting (so mostly empty=poor, mostly full = very good; fill could be non-linear, applied to any scoring method etc; rotational fill vs vertical fill, etc). Circle with F in for fault, instead of filling it based on rating.
- The fill of each blob could also be coloured based on maturity, if wanted.
- after 4+, either just show last four (to allow the table to age),
- Optionally, a 2-figure value in each blob showing year when tasted
I considered various blob types, smiley faces etc - but a page plastered in smiley/unhappy faces seemed tough to read, whereas dense vs less-dense areas of colour are intuitive at a glance. I never did create the table, as ideally would want it to auto-update from all tasting records, and had not got that far through. Figured I'd share the idea as food for thought.

[flash_uk]

Perhaps something like below.
Top left: How many times I have tasted
Top right: How many bottles I have
Bottom left: Status and Year evaluated (in this case, in 2014 I thought "Keep")
Bottom right: Next time to crack one open

In the middle: My rating

[jdaw1]

- Each horizontal line is a house
- Each vertical is a year

That is not the traditional orientation.

[PhilW]

Perhaps a little busy (I'd keep any stock record separate, though, unless this were an automated output) but yep, and assuming the colouring is automatic based on the rating.

- Each horizontal line is a house
- Each vertical is a year

That is not the traditional orientation.

I know; it's just my preference (I generally prefer time on the horizontal axis of a chart).

[flash_uk]

Is it permitted to attach an Excel spreadsheet to a post? If so, I'll attach an unpopulated version of the matrix I have created. It comes with a macro which allows you to select which houses and years get populated onto the template, and for the houses I have started with it also has a template which aims to show which years are declared and which not.

[jdaw1]

Is it permitted to attach an Excel spreadsheet to a post?

No. If you email it to me I could host it on my website.

But that would have a convoluted updated process. You have a website. Is there a reason not to do the obvious?

[flash_uk]

Is it permitted to attach an Excel spreadsheet to a post?

No. If you email it to me I could host it on my website.

But that would have a convoluted updated process. You have a website. Is there a reason not to do the obvious?

Ah of course. That would make sense Will put it up tomorrow - file is on computer in office.

[flash_uk]

This link will take you to the spreadsheet matrix.

There are 3 tabs:
Master Data: lists houses, years, the font for the matrix, and a flag to decide whether or not to colour cells according to years declared. just insert a cell to add a new house, or add a year onto the list for additional years. Place an x next to the houses and years that you wish to appear in the matrix. To colour cells for years which have been declared and not declared leave the Colour Declarations flag as "Yes" otherwise delete the "Yes".

DecMaster: This grid attempts to flag which years have been declared by a house (marked with a D) and which not. The grid only contains certain houses and years, is not complete, and is a very first draft likely to have several errors. As noted above, you don't need to use this if you don't want it. That said, if anyone has a decent list of years declared for houses, if they can share it with me I can update this grid more accurately. One large caveat, with this version, if you add extra houses and years onto the master data tab, and include declaration colouring using this grid, you will need to adapt the declarations grid appropriately to include the additional years and houses otherwise the whole thing will go pear shaped. I'll get round to fixing it so that this works a bit more neatly in future.

Matrix: this tab is blank to start and has the matrix populated at the end.

Once you are all set, then in Excel run the macro called "Populate_Matrix" and sit back for a few minutes. It will likely appear that Excel has frozen for a while, but it should come back to life eventually. I would suggest opening Excel fresh to use this spreadsheet. Don't use it while you are busy working on other spreadsheets just in case...

I did try and make it resize the page to A3 and fit to single page but not sure that has worked...

If you find it doesn't work or something looks wrong, please PM me rather than clogging up the thread with a bunch of issues. Thx.

[Alex Bridgeman]

Interestingly you have approached the problem of how to fill your cellar from the complete opposite of the way I did it.

I looked at the age of the port I like to drink, the number of bottles I pull out of my cellar each year, estimated my drinking life and then modelled how many bottles of each vintage I would need to own / buy to have the cellar I wanted.

Once I knew which vintages (or groups of vintages - my model allowed me to overstock on, say, 1985 and understock on 1984) I needed to buy more of I then kept an eye out on vintages and retailers and bought opportunistically as budget allowed. When I found a particular shipper / vintage combination I'd look it up on the tasting notes or perhaps buy it anyway in order to have variety in my cellar rather than too much focus on any particular shipper - I bought a case of Croft Roeda 1980 not because I expect it to be tremendous port, but because I have little Croft from the 1980s and needed more 1980 vintage.

[flash_uk]

Interestingly you have approached the problem of how to fill your cellar from the complete opposite of the way I did it.

I'm not so sure we have such a different approach! As I have little in the cellar to begin with, I am starting by populating with some no brainers which if consensus on TNs is a fair reflection, will give me pleasurable drinking and some bottles which will make useful contribution to tastings I might have the good fortune to attend. Moving on from there,as I begin to refine what I like, and consider which gaps in the cellar could be filled, I will increasingly be more selective about what I choose to purchase. The only purpose of the matrix is to give some kind of handle on the relative merits of different houses and vintages.