Height/Width: A New Mechanism for Ranking Shows

[Dan Purcell]

It's occurred to me that the way we typically "rate" Phish shows --
assigning each show a Scott Jordan number from one to ten -- isn't
particularly useful. The biggest problem is that the scale is totally
linear; it allows only one measurement of quality. It is very hard to
capture, let alone meaningfully describe, the precise quality of a show
with a single number.

It would make much more sense, I think, to assign each show *two*
rankings, which I call the "height" and "width" ranking. The definition
of these terms should be instantly obvious to everyone; a show's "height"
represents the quality of the show's peaks, while the "width" rating
measures the overall consistency of the show. In other words, a show
like 12/29/94 Providence, with a preternaturally stunning David Bowie and
nothing else, would have a relatively high height but little width.
Conversely, a show such as 11/30/96 Sacramento, which is uniformly well-
played but lacks a showstopper jam, would have substantial width but not
as much height.

Not only are these concepts useful, but I think any honest attempt to
determine which Phish shows are "the greatest" must take account of them.
Think about it: when we say a show was "great," what do we mean, exactly?
Do we mean that *at some point* during the show Phish achieved a truly
fantastic level of improvisational majesty, or do we mean that Phish
sustained a level of excellence over the *whole show*? In order to
meaningfully suggest that a show was "great," you have to answer that
question.

By "height," I don't necessarily mean a show's single peak, like the
Hoodesque jam during the 12/29/94 Bowie or the unplugged section of the
10/7/95 Hood or the thunderous initial moments of the permagroove jam
segment of the 12/9/95 YEM. Instead, I'd define "height" as the average
of the various peaks of a show. By way of example, let's look at
12/30/93. At that show, the obvious Everest is the stunning Mike's ->
Horse, but the show also contains a fantastically energetic (if slightly
hurried) Slave, a terrific Bowie, and the hockey rink/tidal wave
Colonel Forbin's. On the strength of these peaks, and primarily the
Mike's -> Horse, I'd give the show a 9.0 for height.

By "width," I mean consistency throughout the show in every facet. No
botched intros. No irritatingly sloppy harmonies. No obvious fuckups in
the composed sections of Reba or Coil. No blatant miscommunications
between the band members. A discernible absence of tension, whether the
result is the outright silliness of 7/13/94 Big Birch or the remarkably
focused tradeoffs of 12/31/95 MSG, makes a show a prime candidate for
good width. Let's look at 12/30/93 again: the opening Bowie has Trey
teasing Dream On all over the place; the Curtain -> Sample segue is
perfectly timed; the Mockingbird is well-played; the "segue" into
Bathtub Gin clearly shows the band paying attention to each other. In the
second set, Trey leads Mike down a blind alley in Mike's Song, and Mike
heedlessly follows him. I suspect most of you know perfectly well about
this show (PYITE! McGrupp! Purple Rain!), so I'll spare you the details,
but the basic point is clear: the show is focused and consistent from the
opening hi-hat of Bowie to the closing chord of Good Times, and merits a
solid 9.5 for width.

In other words, if you draw a line graph of a Phish show (giving each song
a rating between one and ten, just to simplify), the height of a show
would be the highest point on the graph (or the average of the highest
three or four points). The width of the show would be the total area
under the graph.

In introducing these concepts, I don't mean to suggest that they're
entirely separable. As you'll note, evaluating "the average of the
various peaks of a show" can, if taken to its logical extreme, start to
look like evaluating *the whole* show, and thus height can start to look
like width. At the same time, examining versions of songs for tightness,
energy, and communication can fast degenerate into analysis of the
particular high points of the show, thereby turning width into height.
Despite the possibility of reductio ad absurdum, though, I do think the
concepts are meaningful and useful if employed appropriately.

You hear this argument all the time. Someone will mention a particular
show as great because of two or three moments, and another person will
responding that the show also had a botched Coil, a Rift where the sound
system exploded, or an If I Could that bored the audience half to death
with its utter dullness. Or someone will say that they didn't enjoy a
show all that much (this will usually be Charlie Dirksen) and another
netter will follow up by saying that the energy at the show was great, or
that all the songs were played very well despite the lack of peaks. We
all value different things in shows, and this fact is constantly brought
to the forefront in the world that is rec.music.phish.

Let's look at a few more examples. Which show is better, 8/20/93 Red
Rocks or 6/17/95 Gainesville? If you're talking width, I think 8/20/93
wins easily; the show is one of the most gloriously upbeat and positive
listens in the Phish catalog. But if you're talking height, it's a lot
closer, and I think 6/17/95 beats out the Red Rocks show on the basis of
it's truly stellar Tweezer -> JBG -> Tweezer -> McGrupp, and a
reasonably beautiful Harry Hood.

Or how about this: which show is better, 8/15/93 Louisville or 2/20/93
Atlanta? Again, if you're talking width, I think the Roxy show, with its
uniformly well-played first set and legendarily bizarre (and mostly
inspired) second set, is an easy pick. But if you're looking at height,
the Louisville show is probably better, with the Hose-influenced Stash
and dark, experimental Tweezer, both of which reach peaks of
improvisational bliss untouched by either Roxy set.

(A brief aside: for the baseball fans out there, this is a little like
evaluating which baseball player is best. To begin such an analysis, you
have to ask whether you're looking at the peaks of that players career,
or the total value the player accumulated over the course of his career.
Was Sandy Koufax a better pitcher than Phil Niekro? In terms of peak
value, the answer is obviously yes. But in terms of career value, Niekro
won over 150 more games than did Koufax, so the answer there is no. Was
Mickey Mantle better than Willie Mays? In peak-value terms, sure; in
career-value terms, no. For an articulate discussion of this topic, which
inspired a lot of my ideas here, see Bill James, The Historical Baseball
Abstract at 300-302 (2d ed. 1988).)

I'm sure everyone gets the idea by now, but just to hammer the concepts
into the ground, here are some examples of how I personally would rate
various notable (or not so notable) shows according to this method.

HEIGHT

10 12/31/95 (many stellar jams, including an utterly surprising
Drowned -> Lizards, a Hall of Fame Mike's Song and Weekapaug, a
very good YEM and JBG)
9 12/30/93 (as noted above); 8/13/93 (Gin -> Ya Mar, Mike's ->
Lifeboy)
8 8/15/93 (Stash and Tweezer truly notable for the period, good
Hood); 8/16/96 (only a great DwD, but very good Mike's -> Simple,
Gin, 2001, and Hood)
7 11/22/94 (Funky Bitch into "What the?"); 12/29/94 (a stand-alone
Bowie that's good enough to redeem the rest of this sorry, soggy
show)
6 6/17/95 (as noted above); 6/16/95 (first experimental Runaway Jim,
wonderful Oye Come YEM)
5 10/18/94 (lots of width, though); 4/18/92 (great show, but only
the Linus and Lucy Hood shines abnormally, and an average Phish
show has at least one unforgettable jam)
4 8/20/93 (despite the awesome width, no jam here really stands out)
3 7/8/94 (another fairly wide show, but only YEMFrankYEM soars above
the crowd)
2 9/29/95 (vaguely interesting YEM, no real second set jams)
1 9/27/95 (lame tour-opening effluvium)

WIDTH

10 10/31/94 (near-perfect White Album, showstopping Reba, perfect
Coil, wonderfully subtle and dynamic Slave)
9 8/20/93 (makes me smile like no other Phish show); 12/31/95 (Reba,
Coil, Mockingbird not exactly tight, but forgivable)
8 10/29/94 (consistently brilliant and adventurous; lots of segues);
8/16/96 (remarkable absence of crap)
7 5/28/89 (hilarious, great setlist, very playful)
6 6/9/95 (underrated gem, with terrific Antelope and everything else
more than adequate)
5 12/1/95 (fantastic Halley's -> Mike's -> Weekapaug, but everything
else fairly average, and end of Week is downright depressing);
11/29/96 (worthless first set, good second set, standout Simple ->
Sparks and Taste)
4 11/1/91 (no standout versions, but reasonably tight show)
3 12/3/94 (good Antelope, tight Guyute, little else)
2 12/29/94 (apart from the Bowie, nothing)
1 9/27/95 (I hate to keep picking on this show, but, uh, it wasn't
good)

The above are brief sketches, and are just my opinion anyway, so I'm sure
they contain errors (or represent the addled ravings of a known cretin,
depends on how you interpret them). But whether you agree with the
subjective ratings, you hopefully understand what I'm driving at, and can
apply the methodology yourself if you want. Remember: we don't really
*know* which show is better than any other. I don't, you don't, Charlie
Dirksen doesn't, AJ doesn't, Craig DeLucia doesn't (even though he is the
Phishiest Dankmeister). We all have our opinions, and when you put them
all together you may find that something resembling a consensus develops.
So bear that in mind; I'm not trying to tell you any of my ratings are
definitive.

I do feel, though, that the linear rating system is unhelpful and makes it
hard to really capture what makes a particular Phish show unique. I
realize that adding a second rating may not adequately address that
problem. (We could, for example, use *three* ratings: Hose (free-form
improvisation), Jam (improvisation within structure) and Tightness (lack
of obvious fuckups), or even more, but eventually the data would become
so hard to readily interpret that it'd be useless.) I think bifurcating
the existing system adds more than it detracts, so there it is. Feel
free, of course, to ignore it.

Dan...

[Herschel A Gelman]

Dan Purcell (dpur...@uclink3.berkeley.edu) had written:

| In other words, if you draw a line graph of a Phish show (giving each song
| a rating between one and ten, just to simplify), the height of a show
| would be the highest point on the graph (or the average of the highest
| three or four points). The width of the show would be the total area
| under the graph.

Isn't this the point where you stand on your desk and tell us to rip the
pages out of the book? ;)

| Or someone will say that they didn't enjoy a
| show all that much (this will usually be Charlie Dirksen)

BWAHAHAHAHAAA...

Oh, umm, back to the topic of the post: hmm. Yes. Dan's two-dimensional
rating system is good. I don't know how well it'll catch on, just because
a lot of people might just be too lazy to think about a show enough to
judge both the height and width. I know I probably would be.

The most important thing I learned from this post is "I need to get
11/22/94".

--
Herschel "Newbie Offer: first two people to e-mail me and complete the
quote "addled ravings of a..." can have a copy of 8/14/96
Hershey (mine's DAUD1 from Schoepps mics) for blanks and
postage. Hint: the quote is by Dan Purcell, the guy whose
post I'm replying to. Hint hint..." Gelman

[GMarsh4719]

In the spirit of using Ebonics to teach english, I think Dan's theory
should be
be used to explain Calculus (Icculus Newton.... He wrote the freakin'
book, OK!!!)
to freshman college phishheads. My word, if I had learned Simpson's rule
for
approximation, definate integrals, etc. by analyzing Phish's annual NYE
runs
instead of voltage and current phase lag in inductive circuits, I'd be
much happier
today. :^) Seriously, though, I enjoyed this post immensely, and agree
that the
area under the curve is an ideal way to derive a single, meaningful
numerical index. (although most peoples opinion on any given high point
can differ quite
significantly -- ie. 12/30/96 Guyute.. some loved, some hated)...Anyway,
its back
to the chaukboard for me. Im trying to figure out a way to use Fourrier
Series
to come up with a unique sinusoidal equation for a Phish show based on the
weighted exponential contributions of the four members as <*(((><
approaches
infinity. (just babbling) Peace, Greg Marshall.

[CBertolet]

I like that, Dan. Well supported thinking. I might not use it, if only
because I like words way too much and prefer to ramble and ramble and
prattle and babble endlessly than...

But seriously, this should be useful for more, em, left brained people to
employ in reviews, and for shorter reviews. Good post.

"Democracy is not a static thing. It is an everlasting march." - FDR

chris bertolet

[CBertolet]

[CBertolet]

[CBertolet]

[CBertolet]

[CBertolet]

[CBertolet]

[Dan Hantman]

GREAT IDEAS!!!! All should use these...YEA! Thanks, Dan
--
***************************
* Dan "Helmet Boy" Hantman*
* *
* hel...@mboxes.com *
* *
* A little nonsense *
* now and then *
* is relished by *
* the wisest men. *
* *
***************************
http://www.geocities.com/SoHo/4489/index.html

[Adam Russell]

OK....looking back on the two Spectrum shows from this new angle:

This is all from memory, as I have not yet gotten the tapes:

12/28:
I think the width here is very nice. All the songs were very solid
versions: The Runaway Jim was not very expansive...but it was nice
smooth excellent jamming, albeit short. NICU, although Trey flubbed a
note, was solid, with some nice work by Page. Wolfman's Brother was very
nice jamming, one of the best versions I've heard. Ginseng: solid,
SOAM....very, very nice version...this was a definite highlight.
Makisupa into Maze was very nice...the maze was a standard one...nice
powerful jam. The Delay loop>TMWSIY was absolutely gorgeous...pretty
work here by Trey, astounding musically, although I don't think it was
as good as the 12/31/95 Mike's one which is one of my favorite moments
of phish in my entire tape collection. The Mike's was spacier than
usual, with what many people think was a Simple jam (I thought they just
modulated into a new key, but I haven't heard the tapes) The Weekapaugh
was a little below average, with the exception of the absolute great
ending solo by Page. Encore JBGoode, I thought was better than New
Year's--If anyone read the Philadelphia Inquirer review of the show,
the entire artice focused on this song, and how good it was...sorry I
didn't type it up:(
Looking back on the whole show, the width would be an easy 7 or 7.5 Not
many shows have this depth of super-solid songs. Although none of the
jams were particularily mind-blowing, every song was very strong. The
first set was one of the better ones that I've seen.
However, the height is below average--maybe a 4.0. The only great jams
were SOAMelt and the DDJ>TMWSIY. Less truly great moments than a typical
show.
However, I do whole-heartedly disagree with people who say that this
show was a typical 5.0 great phish show. Most shows don't have this
amount of super-solid songs, where every song is powerful. Especially
considering that the first set when compared to other _first_ sets is
clearly much better. I would give the show an above-average rating just
for its width. Although I do really enjoy those truly deep and jammy
moments of this show and this one might be slightly lacking, I would
easily rate the show an above average 6.5 Just place this show on a par
with all other shows since say 1992. It's better than most of them, I
would say.

12/29:
The width was pretty poor here. The first set was very weak...the nice
suprise of a caravan did not make up for the Cavern, Guelah, Rift which
were all pretty much heartless. The Taste was much better than average
tastes, and I hope the boys work on this one, but this being said...the
superb outro jamming was only a minute or two. The Free was _very_ below
average...probably a 2.0 on the Free scale. As soon as the composed
singing section was over, Trey went right for the drums, the jam went on
for about 5 minutes, going nowhere, Trey returned to guitar, and started
singing the end of the song. When there is an entire set that is this
lackluster, I don't think there can be much width. The width for second
set is 9.5, as one of the best widths ever: the entire set was
absolutely stellar in terms of entertainment, and nice jamming.
The depth also occurs wholly in the second set. With the _nice_ Bowie,
the best Day in the Life I've heard, the oustanding Gin>Lizards, as
detailed by Benjy, and the wacky YEM and Harpua. The height for this set
has to be a 7.5, as one of the best ever, not for mind-blowing
jams(Providence Bowie) But for the insanity. YEM wasn't musically
awe-inspiring, but it was very fun. This set is very hard to rate, based
on its wierd characteristics. I don't think the show as a whole can be
put as a single value, I would think that each set would have to be done
seperately, as they were night and day.

Recap:
12/28 Width: 7.5
Height: 4.0
Show: 6.5

12/29 1st Set: Width: 2.5
Height:2.0
Overall: 2.0
2nd Set: Width: 9.5
Height:7.5
Overall: 8.5

Thanks for bearing with this rambling post and all the silly numbers
that can't possibly describe music. Great idea for the new scale, I say:
use it.

Giving myself a 5.0 standard performance for this rating,
Adam
http://www.personal.psu.edu/users/adr128

[Dan Purcell]

Herschel "Cheddar" Gelman wrote in response to me:

: | In other words, if you draw a line graph of a Phish show (giving each song

: | a rating between one and ten, just to simplify), the height of a show
: | would be the highest point on the graph (or the average of the highest
: | three or four points). The width of the show would be the total area
: | under the graph.

: Isn't this the point where you stand on your desk and tell us to rip the

: pages out of the book? ;)

O CAPTAIN MY CAPTAIN!!!!!

I should probably apologize for the schoolmarmy tone of the paragraph that
Herschel quotes above. I don't mean to suggest that anyone should really
go out and *draw a line graph* of a Phish show, just that if you did, you
could determine height and width by the method I suggest.

And besides, I would rather see some impressionistic writing about Phish's
music than this verkakte analytical crap I usually churn out. Analysis is
certainly easier than impressionism. But given that most people post show
reviews in terms of "that show was a 6 out of 10," I figured I might as
well try to improve *that particular rating system*. But obviously a
number, or two numbers, could never capture the essence of any Phish show,
or indeed any piece of music, or indeed any work of art.

: | Or someone will say that they didn't enjoy a

: | show all that much (this will usually be Charlie Dirksen)

: BWAHAHAHAHAAA...

You know the rule, Herschel: if you drop Charlie's name in your posts,
he'll BE YOUR FRIEND. And then he'll GIVE YOU TAPES.

: Oh, umm, back to the topic of the post: hmm. Yes. Dan's two-dimensional

: rating system is good. I don't know how well it'll catch on, just because
: a lot of people might just be too lazy to think about a show enough to
: judge both the height and width. I know I probably would be.

I don't know that it takes a lot of effort. But then again, I'm normally
too lazy to stand upright, so what do I know?

: The most important thing I learned from this post is "I need to get
: 11/22/94".

Yes, you do. As do we all. This largely unknown show took place in my
home town of Columbia, MO one day before I returned home for Thanksgiving.
I missed the show, but my mother (!) and cousin were in attendance. Mom
reported "a lot of people smoking pot," but also noted that Phish "could
really play!" The second set (are there even tapes of the first set?)
opens with a Funky Bitch that segues magnificently into a wierd-ass jam
that's initially very groovy but turns odd, spacy, and beautiful rather
quickly. Finally we get a segue in JCOG, then Cry Baby Cry, then the
never since-repeated combination of Curtain -> Blackbird. An energetic
Runaway Jim segues smoothly into BBFCFM; the boys put down their
instruments during a break in the song, grab their acoustics (Trey on
guitar, Mike on banjo, Page on standup bass, Fish on mandolin or
washboard) and finish the song acoustically. After the acoustic set, the
show closes with Harry Hood and Highway to Hell, Lizards encore. Not a
bad evening, all in all.

It ain't easy livin' free,

Dan...

[Mr. Mood]

In article <5bicn4$1...@agate.berkeley.edu>, dpur...@uclink3.berkeley.edu (Dan Purcell) wrote:
>Herschel "Cheddar" Gelman wrote in response to me:
>

>: | In other words, if you draw a line graph of a Phish show (giving each song

>: | a rating between one and ten, just to simplify), the height of a show
>: | would be the highest point on the graph (or the average of the highest
>: | three or four points). The width of the show would be the total area
>: | under the graph.
>

>: Isn't this the point where you stand on your desk and tell us to rip the
>: pages out of the book? ;)
>
>O CAPTAIN MY CAPTAIN!!!!!
>
>I should probably apologize for the schoolmarmy tone of the paragraph that
>Herschel quotes above. I don't mean to suggest that anyone should really
>go out and *draw a line graph* of a Phish show, just that if you did, you
>could determine height and width by the method I suggest.
>

[snip]>
>
> Dan...

Dan, I was gonna make the same references to Dead Poets', but you beat me to
it!!

Let's totally SQUEEZE the life and enjoyment out of Phish, shall we?! Nothing
like a mathematical formula to get the blood flowing in the morning.

mood

[CBertolet]