2011 Horizontal - London, Sunday 23rd April 2051

[jdaw1]

Just gone from a Fonseca LBV to a Taylor's; is that not a demotion?

See Cruz Ruby Revisited.

[jdaw1]

Julian, the organisation of this tasting is a mess, can you please sort it out? Hurry along man, there is no time to lose!

Lordy: is it time to take the food order? Should I assume that we will all still have teeth?

From The Code of Health and Longevity (1818), by Sir John Sinclair, page 123:

It is a custom, which almost universally prevails in the northern parts of Europe, to present a dram, or glass of liqueur, before sitting down to dinner. It answers the double purpose of a whet to the appetite, and an announcement that dinner is on the point of being served up. As the practice has continued so long, most probably it has been found to answer the first of these objects, or at least to do no harm; and the other has the convenience attached to it, of letting those of the company engaged at cards or billiards know, that they should stop, without beginning a new game or party. Along with the dram, is presented on a waiter, little square pieces of cheese, slices of cold tongue, and dried toast, accompanied with fresh caviar, &c.

A celebrated physician, when far advanced in life, found a similar stimulus necessary. After his dinner, which consisted chiefly of broths and fish, (as he could not eat animal food from want of teeth), he drank a pint of port daily, and concluded with a wine glass filled with white sugar and as much rum as it could hold; so that he had fully half a glass of spirit. The same dose was taken every morning about twelve o’clock; and if visiting at that hour, he was accustomed to call for it, as regularly as when he was at home. It was never understood that his health suffered from that practice.

[DRT]

Julian, the organisation of this tasting is a mess, can you please sort it out? Hurry along man, there is no time to lose!

Lordy: is it time to take the food order? Should I assume that we will all still have teeth?

Starter:Soup
Main: Soup
Savoury: Soup
Dessert: Custard

[DRT]

Are we reconfirming again? I reconfirm!

I reconfirm.

[Andy Velebil]

Are we reconfirming again? I reconfirm!

I reconfirm.

As do I.

[PhilW]

I'm not dead yet. I reconfirm.

[differentdave]

I can reconfirm both my and my sons attendance. Have been itching to taste a few of the 11's this year but must show restraint. I will be celebrating my 80th birthday shortly after this offline -- My father lived to 82 and was enjoying wine until the end, I am pumped!

[jdaw1]

I can reconfirm both my and my sons attendance. Have been itching to taste a few of the 11's this year but must show restraint. I will be celebrating my 80th birthday shortly after this offline -- My father lived to 82 and was enjoying wine until the end, I am pumped!

Most all of us have offspring. If signing on children were allowed, there would be many dozens of my-daddy-signed-me-up-for-this. A policy decision has been taken that this isn’t fair (DRT ruling). The seats belong to Port enthusiasts, giving some amount of priority to those who signed early, and some amount of priority to those with bottles. Offspring who are genuinely enthusiastic about Port will have to charm their own way into the event.

[Justin K]

Julian, not wishing to sart an argument with you regarding non attendance of offspring but you're being a bit hasty. Wouldn't it be good idea to have someone to catch the dribbles and generally keep an eye on us (medical training not obligatory but a definite plus)? I'm also a bit worried about those narrow stairs, I don't want to risk my hip and knee replacement replacements; a young strong she, he or it would come in handy. I think Derek may have to look into this.

[PhilW]

Julian, not wishing to sart an argument with you regarding non attendance of offspring but you're being a bit hasty. Wouldn't it be good idea to have someone to catch the dribbles and generally keep an eye on us (medical training not obligatory but a definite plus)? I'm also a bit worried about those narrow stairs, I don't want to risk my hip and knee replacement replacements; a young strong she, he or it would come in handy. I think Derek may have to look into this.

I believe the organiser already authorised the presence of a nurse; Julian's objection is I think to someone's suggestion of seats passing by inheritance instead of to TPF reserve list, rather than to preclude assistance.

[MigSU]

Offspring as a +1: yes.
Offspring as a replacement: no.

[jdaw1]

Offspring as a +1: yes.

So bring one of mine as my +1, who thereby takes the seat of a genuine Port enthusiast? No.

[Glenn E.]

My reading of the rules:

Anyone who wishes to attend and drink (hereinafter a "drinker") must have their own account here on TPF and sign up separately. Accommodations for industry guests to be arranged by the organizer (DRT) or his duly appointed representative (jdaw1).

Others may attend as aids to the drinkers, but do not receive their own portions, and their attendance is subject to the capacity of the event room once that has been determined. Whether or not their drinker shares a sip with them is entirely at the drinker's prerogative.

[MigSU]

Offspring as a +1: yes.

So bring one of mine as my +1, who thereby takes the seat of a genuine Port enthusiast? No.

Well, I was operating under the assumption that the offspring was also a Port enthusiast (but not a member of TPF, so can't sign themselves up for this). Otherwise they'd have no interest, and would only be there as a care worker (in which case they could be excluded from imbibing any Port at all).

[jdaw1]

This tasting was started with the default process for handling people. That process copes well with the then-usual outcome, in which about fourteen would want to attend, a short waitlist holding the others. It has become clear that such a process isn’t robust to a tasting being announced 37¾ years before the event.

As that process doesn’t work, I will soon be announcing a change. This is its proto-announcement, in that it can be changed in the light of reasoning received.

---

• The possible attendees are being combined into a single list.
• Nearer the time, the organiser (originally Derek, now me, but not necessarily remaining me) will select the fourteen attendees. In making that selection, the organiser will have regard to the following.
  • First: Bottles that person would bring.
  • Second: Enthusiasm for Port, at least partly measured by scale and quality of role and presence, both offline and online.
  • Third: pleasurability of company.
  • Distant fourth: timeliness of requesting a place. (This isn’t quite the same as ordering by date of first request: somebody now young who becomes a Port enthusiast might be very timely even if making the request a decade from now.)

---

Some people hitherto on the list won’t attend by reasons of death or incapability. Others might not attend because the organiser has chosen an attendee thought more suitable. Doubtless those losing by the change won’t like it. Even so, please recognise its broader fairness, and play nice.

— JDAW (new organiser) and DRT (founder organiser).

[DRT]

This tasting was started with the default process for handling people. That process copes well with the then-usual outcome, in which about fourteen would want to attend, a short waitlist holding the others. It has become clear that such a process isn’t robust to a tasting being announced 37¾ years before the event.

As that process doesn’t work, I will soon be announcing a change. This is its proto-announcement, in that it can be changed in the light of reasoning received.

---

• The possible attendees are being combined into a single list.
• Nearer the time, the organiser (originally Derek, now me, but not necessarily remaining me) will select the fourteen attendees. In making that selection, the organiser will have regard to the following.
  • First: Bottles that person would bring.
  • Second: Enthusiasm for Port, at least partly measured by scale and quality of role and presence, both offline and online.
  • Third: pleasurability of company.
  • Distant fourth: timeliness of requesting a place. (This isn’t quite the same as ordering by date of first request: somebody now young who becomes a Port enthusiast might be very timely even if making the request a decade from now.)

---

Some people hitherto on the list won’t attend by reasons of death or incapability. Others might not attend because the organiser has chosen an attendee thought more suitable. Doubtless those losing by the change won’t like it. Even so, please recognise its broader fairness, and play nice.

— JDAW (new organiser) and DRT (founder organiser).

Agreed.

[jdaw1]

This is its proto-announcement, in that it can be changed in the light of reasoning received.

No objections have been received, Indeed, no reasoning at all has been received. Therefore this is henceforth de-proto’d.

[jdaw1]

The Rules

• The possible attendees have been combined into a single list.
• Nearer the time, the organiser (originally Derek, now JDAW, but not necessarily remaining JDAW) will select the fourteen attendees. In making that selection, the organiser will have regard to the following.
  • First: bottles that person would bring.
  • Second: enthusiasm for Port, at least partly measured by scale and quality of role and presence, both offline and online.
  • Third: pleasurability of company.
  • Distant fourth: timeliness of requesting a place. (This isn’t quite the same as ordering by date of first request: somebody now young who becomes a Port enthusiast might be very timely even if making the request a decade from now.)

[Doggett]

I was on the verge of moving to a plant based (grape) vegan diet, joining a yoga community and meditating for the next 30 years in the hope of being alive for this tasting, but with no guarantees of a place… bugger that and I shall carry on as was!

[jdaw1]

no guarantees of a place

The best guarantee available is …

Magnum of 2011 Noval Nacional

[Glenn E.]

no guarantees of a place

The best guarantee available is …

Magnum of 2011 Noval Nacional

Hopefully followed closely by all 3 super cuvées!

[jdaw1]

The Society's Exhibition Vintage Port 2011:

'Absolutely singing' now, this Port from Symington Family Estates, put together exclusively for members, comes from a great vintage and our Wine Champions tasters voted it a 2025 winner. ‘Robust and rich’ with ‘lovely texture’, its rose-petal aromas and concentrated, structured palate proved irresistible in our tasting. Your cheeseboard deserves it!

Price:£43.00 Bottle

Price:£258.00 Case of 6

Have we knowledge of the Wine Society’s 2011? Is it a custom blend? I.e., not the same as any of G D W V Ck Rr SW QH, nor Waitrose nor BBR.



Edit: AHB’s list, in this thread, of 2011s describes it as “Special SFE blend for The Wine Society”.

[NigelM]

I'm available. No seating needed, just a spot on the table for my Urn. (Simply add port)

[winesecretary]

Please add me to the reserves for this one. I’ll only be 78. I have Cockburn, Croft, Fonseca and M&S.

[jdaw1]

Per the most recent rules, there has not yet been a decision about who will attend. That will come later.

[mcoulson]

Can I be added to the hoping to attend list please

[akzy]

So after enjoying a Rose port from my Quevedo advent calendar I started reading this...

[*]Second: Enthusiasm for Port, at least partly measured by scale and quality of role and presence, both offline and online.
[*]Third: pleasurability of company.

I became terribly concerned. How would one actually objectively decide this?

Then I saw the below

Some people hitherto on the list won’t attend by reasons of death or incapability.

Would we reasonably have to invoke these powers?

So naturally I did what any sane TPF member would do and spawned 1,000,000 universes to see what happened.

TL;DR There's about 15% chance that the current list would be full for a seating of 26 people (yes I know this is larger than the 14 said but the 1970 had 26 and the numbers are more interesting) with an average of 23 people remaining for the 30 signed up. This also means only a 15% chance on relying on Julian's judge of character.

Making a stereotypical TPF

I started off by creating a random sample of The Port Forum members for each universe. This is a normal distribution with a mean age of 55 with a standard deviation of 10 years. This means that 63% of members fit between 45 and 65. I chose 55 because I have five fingers on my left hand so it has sentimental meaning to me.
Next I had to decide on the biological sex. I set this at about 10:1. 10 was chosen because I have 10 fingers.

Running a single universe.
Who is going to die? Fortunately, there's a big data set on survival rates in the UK, notably "lx" data which is if we started with 100000 people, at a certain age, how many of those are left? This is sorted by both male and female. I'm using the cohort data which presumes healthcare gets better and you live longer YMMV. Using this, I can calculate the odds of someone surviving to 2051 based on their age. I then compare this to a random number to determine if that participant lives or dies. This is run for all members returning how many people survived from that universe.

Running multiple universes
Rinse and repeat the above a million times. Also known as the Monte Carlo method. Using this we can just see what the dice would have said on average.

The results
For a starting pool (N=30) and a target seating (K=26) and time to event (T=26) and a demographic of 10:1 men with average age 55 and std deviation of 10 years, there's only a 15% chance we'll have a full tasting. The average attendance would be 23 (.3 - unsure what this would look like) people.

tpf_2051_tasting_survival.png (271.12 KiB) Viewed 261 times

If we instead held the Mumsnet 2011 vintage port tasting (average age ~42 with std deviation of 8 year, 19:1 female/male ratio - courtesy of Gemini), there's a 99.4% chance of a full tasting and Julian would have to cull the others.

tpf_2051_tasting_survival_mumnet.png (272.42 KiB) Viewed 261 times

What I'd like to do but I'm lacking data
If we can also eliminate people who are in long term care, that would help. Unfortunately I can't find a good data set to separate this out from dead people.

If all attendees could please provide

• Current age
• Biological sex
• Long 16 digit on card
• Have they even been a smoker?
• CV2 number from card
• Mother's maiden name

Then I can create a more accurate model and also purchase more heavy AI analysis.

[hadge]

I think that you are missing how much port do we consume in a year and what the impact of this is on our bodies.

[winesecretary]

I think that Zak’s post has to go straight into the pantheon of all time greats. Best laugh I’ve had since the Budget and (assuming he only needs the 3,5,6 data points for statistical purposes) a whole lot cheaper.

[akzy]

I think that you are missing how much port do we consume in a year and what the impact of this is on our bodies.

Eurgh. Back to the data mines I go.

I think that Zak’s post has to go straight into the pantheon of all time greats. Best laugh I’ve had since the Budget and (assuming he only needs the 3,5,6 data points for statistical purposes) a whole lot cheaper.

The things Rose Port does to a man.

[Alex Bridgeman]

That was a fabulous post. Zak - thank you for going down a a most entertaining rabbit hole.

Note to self: ask Oscar to put many more bottles of Rosé Port in next year’s advent calendar.

[mcoulson]

Excellent post Zak ... Very entertaining and worth your time doing it .... I note you appear to be assuming that Julian will be there to apply the rules ... Note sure what our policy should be for stand in people !

I could also change my will to donate bottles to the event if I don't make it !

[PhilW]

The things Rose Port does to a man.

Indeed; I was going to make some kind of comment along the lines of "that'll teach us not to drink Pink Port" execpt I enjoyed the post too much. That said I'm going with a much simpler statistical-lies approach of drawing a curve of TPF attendee deaths, which so far is (appropriately) a flat line, and projecting that trend into the future I predict that we'll ALL be able to attend, denial not just being a river in Egypt.

[akzy]

Eurgh. Back to the data mines I go.

Integrating the fact we drink more
Previously we had the 'lx' table which starts with 100k people and each year calculates how many are left. If we have the mortality multiplier (otherwise known as a HR = hazard ratio) which can get the change in probability over a standard year that someone would die due to the effect.

An example, for a standard person (HR = 1), in year one there is a 100k people, for year 2 there is 90k people (i.e. 10k lost). For a HR =0.75, the adjusted year 2 would be 92.5K (100k-10k*0.75). For a HR = 2, the adjusted year 2 would be 80K (100k -10k*2). These aren't the real numbers - just an example.

The hardest part of this is getting reliable data for HR. Literally all over the shop. There's some "leading" UK data that drinking crazy amount of booze still gives a HR<1 which means you'd live longer. The best data I could find was from https://www.jacc.org/doi/10.1016/j.jacc.2017.06.054 which is for the US. It lines up with what I've seen which is a certain amount of alcohol actually improves life expectancy before it turns. See the below image.

alcohol_mortality.png (676.47 KiB) Viewed 111 times

For a male currently aged 55 and HR of (0.75, 1, 1.25, 1.75) the probability to make it to the tasting is (89%,85%,82%,75%).


Updating the stereotypical TPF

Now given we're in Monte Carlo and have our finest Tux on, let's keep rolling the dice. Going back to making a stereotypical TPF, I roll the dice for each participant to see if they drink 0, 0.5, 1, 1.5, >2 bottles of Port per week equating to HRs of (1, 0.75, 0.75, 1.25, 1.75) respectively. I also weighted it such that we're skewed to the higher side and that no one drinks none. I use therefore weights of (0, 0.5, 2, 2.5, 1).

Re-running parallel universes
So the problem is, with all these random variables, the error on the final number is also much more variant. To reduce this error, you have to run more simulations. Now running with 10million universes.

(In a further rabbit hole, this has now necessitated parallel compute because 10million universes would take about 1hour to simulate in my terrible python code. On 20 cores, I get about a 5x improvement in time vs single core. Given that each task is so small, I'll take that as a solid win for the overhead of pooling).

New results

Perhaps surprisingly, for these demographics, the odds slightly improve because of the moderate drinkers low HR. 27% chance of having enough people.

adjusted_tpf_2051_tasting_survival.png (270.32 KiB) Viewed 111 times

Re-running for baseline again but with larger simulation i.e. weightings of (1,0,0,0,0), we get a more accurate number for the original scenario in the previous post of 29.75%. The eagle eyed among you will realise this is me admitting a mistake in my previous code.

It's only when you change the HR weightings to something like (0,0,1,3,2) do you get a chance of 17.7%.

"But what about the Mumsnet 2011 vintage Port tasting?". I'm glad you asked. I gave them weight of [1,2,3,1,0.3] and the odds are now 99.5% i.e. basically unchanged. This makes sense given that it was so high before.

"Zak, what are you drinking tonight?". I was tempted to try hunt the other Rose Quevedo but that's not in the advent spirit. I have a 10YO and Quevedo White for this evening hence a far more sensible update.

[winesecretary]

I am almost weeping with laughter. Yet again the combination of haut-mathematics* and port is providing a new sort of fun.

* I accept that proper mathematicians may point out this is mainly statistics

[jdaw1]

I am very fond of a chart where the x axis is labelled “Drinks/Week”, but does not specify how many magnums decanters are deemed to equal one drink.

[akzy]

I am very fond of a chart where the x axis is labelled “Drinks/Week”, but does not specify how many magnums decanters are deemed to equal one drink.

It took a while to find it in the paper. A drink is 5 fl oz. Of wine (12% abv). I approximated a bottle of port to be 10 drinks.

[Glenn E.]

I am very fond of a chart where the x axis is labelled “Drinks/Week”, but does not specify how many magnums decanters are deemed to equal one drink.

It took a while to find it in the paper. A drink is 5 fl oz. Of wine (12% abv). I approximated a bottle of port to be 10 drinks.

I was also amused by the labeling.

For many years I have also been using 5 wine / 10 Port per bottle as the number of available servings.

"Bottle" intentionally left undefined for Julian's benefit.