Calculating Bourbon Cost vs. Value

[kickert]

I have been thinking a lot lately about values in bourbons. I know what I consider to be a good value (Benchmark, Rittenhouse BIB, Weller SR, etc.) but I wanted a more scientific way of calculating things. I keep a pretty detailed set of tasting notes complete with price and ratings so I figured this should be a pretty easy exercise. I found simply dividing rating by cost did not work because it put too much emphasis on having a low cost. There are plenty of great values in the $20-30 range.

Here was my basic assumption: The cheapest bourbon should have the lowest cost and vice versa, the most expensive bourbon should have the highest rating. Any diviation from that would affect the value of a bourbon.

Before we get into the math, let me give you a key to my formulas:

P = Price of bourbon
Ph = Price of highest cost bourbon
Pl = Price of lowest cost bourbon
Pv = Price
R = "Magic" Ratio
Se = Expected Score
Sh = Highest score
Sl = Lowest score
Sa = Actual Score
V = Value rating
Vt = Value Threshold

First we must be able to calculate the expected score for each price. In order to do this, we must come up with a formula that gives the lowest possible score to the lowest cost and the highest possible score to the highest price. Here is the formula I used:

Se = (P-Pl)R+Sl

Basically what the formula above does is assigns the lowest score (Sl) to the lowest cost bourbon. However, in order to make it work correctly, we need to figure out the "Magic" ratio that results in a perfect score (Sh) to the most expensive bourbon. I used the following formula to figure that out:

Sh = (Ph-Pl)R+Sl

That simplifies to this:

R = (Sh-Sl) / (Ph-Pl)

I calculated two ratios. The first was based on my notes where the highest price (Ph) was $100 and the lowest price (Pl) was $8. My highest score (Sh) was 9.75 and my lowest score (Sl) was 5.5. The "personalized ratio" was ~0.0462. I then did an overall rating where I looked at the highest price bourbon commonly available in my area (PVW23 @ $250) and the lowest price (Ten High @ $6). I also figured the highest score would be a perfect 10 and the lowest score on my scale would be a 5. The resulting "overall ratio" was ~0.0205.

Armed with this information, I am able to calculate the value of each bourbon. This is basically the actual score (Sa) dividing by the expected score (Se). This requires the following formula:

V = Sa / ((P-Pl)R+Sl)

When I ran the numbers for all 55 whiskies in my tasting notes, I found the results it returned were very close to my own estimates. In other words, excellent bourbons at all price ranges rose to the top. A decent bourbon at a cheap price was deemed to be a better value than a good bourbon at an average price.

When using my "personal ratio" here is my top 10 list:

WL Weller 12
Rittenhouse Rye BIB
Elmer T Lee
Four Roses Small Batch
Very Old Barton 86
Elijah Craig 12
Old Charter 10
Evan Williams 1783 No.10
Eagle Rare 10
Buffalo Trace

When using the "overall ratio" here is my top 10 list:

WL Weller 12
Elmer T Lee
Rittenhouse Rye BIB
Four Roses Small Batch
Eagle Rare 10
Very Old Barton 86
Elijah Craig 12
Van Winkle Family Reserve Lot B
Evan Williams Single Barrel Vintage (98)
Old Charter 10

What I found is that if you can afford more expensive bourbon (i.e. look at the overall ratio) then it is easier to recognize values in the higher price range. If you max out at $30 the values are going to be lower priced. This is easiest to observe by looking at the movement of particular bourbons on my list (outside the top 10 list). In light of the bourbons I can afford, GTS is a decent value, but in light of everything out there it is a pretty good value. Likewise, in light of the bourbons I can afford, Benchmark is a pretty good value, but in light of every bourbon out there, it is only a decent value. So then, a good value is really dependant on how much you are able to spend.

As I was writing this tome on bourbon values, I realized I could calculate the threshold at which a bourbon is a good deal. Basically I wanted to find the price point where a particular bourbon would enter the 75th percentile of bourbon values (top 25&#37. To do this, I took the same formula as above and set the value threshold (Vt) where a bourbon entered the top 25% and then calculated that value price (Pv). Here is the formula

Vt = Sa / ((Pv-Pl)R+Sl)

When solved for Pv the result is:

Pv = (Sa + Vt*Pl*R – Vt*Sl) / Vt*R

Of course, you could set your threshold at any level you want. Perhaps you think anything in the top 50th percentile is a value, or maybe you are extrememly value conscious and are only interested in what would make something a top 10 value. Looking at my results I found these stats interesting:

Based on my personal ratio value threshold, WL Weller 12 would still be a value at $49 while Woodford Reserve would have to be priced at under $8 to be a value. Pappy VW 20 would have to be under $30 to be a value, but based on my preferences, I would pay $45 for VWFR 12 Lot B and still consider it a value. At the far end of the spectrum, someone would have to include $13 with a bottle of Early Times for it to be a value on my scale!!

This value threshold information would be very useful when trying to figure out how much a person should spend on dusties. For instance, the $20 I spent on OC12 was appropriate, but if I can only find it for $25 it is probably not worth it.


So what do you all think? My wife is convinced I have spent way too much time working all this out, but I think it is well worth it as I can better recognize "value" bourbons. If you are interested, you can see my excel file here: http://spreadsheets.google.com/ccc?k...fZb-U_cFBLaoMQ

[spun_cookie]

interestign results. You need to continue to update the sheet every 6 to 12 months and then recompair...

As an Engineer... I love seeing equations...

[kickert]

interestign results. You need to continue to update the sheet every 6 to 12 months and then recompair...

As an Engineer... I love seeing equations...

I am still up in the air as to which ratio is most helpful. Once I have that figured out, I will probably include this in my regular notes. The only catch is there are two of us who share the excel sheet with the notes and we each have different ratings. If I use "personal ratios" and "overall ratios" to calculate value and price threshold for both of us we end up with quite a few variables.

[ACDetroit]

Another Engineer piping in! Great idea and I really like the idea! Keep up the good work!

Tony

[chilidawg7]

Not an engineer, but a math teacher (Pre-Cal and Algebra II this year.)

I enjoyed reading your post and can hear my wife (and students) saying the same thing if I were to present it to them: "You have too much time on your hands," which is the common refrain when someone is just jealous that they didn't think of it first.

[Special Reserve]

I love your equations, it reminds me of the p chem days, it made me think.

Wait, I though bourbon was suppose to be fun... Well this is fun. Another member who like GTS 08 imagine that. Make we want to go out and buy some more if I can find it.

[bvscfanatic]

Hmmm, I don't know. It seems to me that price needs to be weighted for proof. In otherwords, dollars per alcohol by volume. It's one of the reasons why OWA is such a great value at 107 proof.

For example, let us suppose that one likes both Bulleit and OGD 114 equally well from the perspective of personal taste (Ben is rolling his eyes at this point, but no matter). In this case, I believe the Bulleit to be slightly more expensive, but let's call them even for this argument. Bulleit is 90 proof, while the OGD is 114. This factor should (it seems to me) be included in weighing the relative value of OGD. Wouldn't you agree?

[jimibourbonhammered]

I was thinking the same thing. I mean I can figure this off the top of my head from notes on 50 bottles.. $25 range = Old Weller Antique 107, $50 range = Pappy Van Winkle 15yo, $50+ range = George T Stagg. Why so complicated?

[kickert]

I can see your point about proof. However, I tend to think it will all come out in the wash when you consider score is about more than just taste. If you like two bourbon's the same but also like a higher proof, wouldn't you give the higher proof a higher score? You could also run into problems with bad bourbons at high proof. For instance, I hate OGD BIB. It does not become a better value just because it has more alcohol in it.

GTS becomes an interesting situation, because if you water it down to ER17 proof and like the flavor of both equally, you get significantly more bourbon of the same quality for cheaper. In that case I would just give a higher score to the GTS.

[Dramiel McHinson]

Slap me naked and hide my clothes!

I really enjoyed this Kickert. It has great potential and could develop into an actual ISO quality standard with the formula being the guiding source of the regulating body... Which should be a panel of bourbon drinkers selected from SB.

I do similar exercises in systems engineering for DoD. Often the perfect solution for a weapon system is too expensive, the technology to immature and the time from concept to deployment too long. In systems engineering we consider a well defined, clear and concise requirement to be the foundation of building a successful system. Our enemy is the interfaces and environment our system will operate in. Sometimes the best result for an affordable price deployed quickly is good enough. Anyone familiar with DoD knows that it's really hard to get it right. But that's another story for another thread.

Let's entertain, for a moment, a system of systems approach to "bang for the buck" bourbon. First we need a working integrated product team (WIPT) made up of all interested members of SB (with time on their hands). Next we need a team lead, perhaps Kickert would like this position. If we agree, we can then begin the process of defining our requirements against our model presented by Kickert.

Our process would go something like this;

1. System Requirements definition/development

2. Functional analysis and requirements allocation

3. Design integration and testing

4. Final design

5. Limited deployment and evaluation (key members actually use the model for decisions on purchases)

6. Full scale use with spiral upgrades as lessons are learned

7. Gather for a party and celebrate solving one of man's greatest challenges, how to buy superb bourbon for a very resonable price.

To make this work, we may have to buy lots of bourbon, drink it, and record our results. Testing our bourbon against key performance parameters, measures of performance and effectiveness as outlined in our requirements document will require dedication to drinking bourbon of all cost points and taste profiles.

We must exercise cost and schedule control which will require careful oversite. I don't know about you all, but my wife has been known to give me the, "who farted," look when I get carried away with a new purchase.

I think a WIPT annual in process review (IPR) meeting at the Kentucky Bourbon Festival to review results and set goals, objectives, and schedule for the next year would be good. Other IPR's at other locations would not be out of the norm for such an undertaking.

Are we up for this?

(we at least need volunteers for testing)