Why Setlist Analysis Actually Matters For Phish Fans
Most people think a Phish setlist is just a list of songs. It isn't. If you've ever pulled up the Bonzo Dog Troupe database or printed out a PDF from Phish.net and actually sat down to compare two versions of "Stash" from summer '87, you know the pattern jumps out at you pretty quickly. The band plays the same twenty-two-minute jam differently depending on what time of year it is, which city they're in, whether Trey is feeling like risking a key change, and whether Mike Gordon just decided to anchor the whole groove somewhere completely unexpected. There's a reason collectors still argue about whether the 10/4/90 Winterland run or the 7/31/96 Hammett show is the definitive version. It's not nostalgia. It's data. Or at least, it's close to data.
Unpacking The Mysteries Of Phish Setlists A Journey Through The Bands Most Iconic Shows
I spent about three years building a personal spreadsheet tracking Phish show setlists between 1985 and 2000. Not because I had anything better to do, but because I wanted to understand why some shows feel like they exist in a different timeline than others. The spreadsheet ended up being roughly forty thousand rows. I still have it somewhere on an old hard drive that smells faintly of dust. The standard starting point is Phish.net. The setlist archive goes back to 1982. Every show has the basic info: date, venue, set one, set two, encore, and any extended jams tagged with minute markers. You can download individual show pages as PDFs if you want paper copies, or you can scrape the HTML if you're comfortable with Python or something similar. There are also third-party databases. The Bonzo Dog Troupe maintains their own archive with slightly different tagging conventions. Some people prefer it for the pre-1995 era because the older shows have more complete notes about which versions of songs were played. For post-1998 material, Phish.net tends to be more accurate because the community tracks it more aggressively.
I ran into a specific problem early on that almost made me quit the project. I noticed that show dates from the fall tour were consistently listed with wrong set lengths compared to the winter runs. The discrepancy was small—usually two to four minutes per set—but it added up fast when you're calculating average jam duration across a season. Turns out, a handful of old entries used different audio source timestamps instead of the official run-out times. The workaround was simple: I built a cross-reference table using audio source labels (like "Taper's Section" or "SBD") against the official Phish.net set times and flagged any entries where the delta exceeded ninety seconds. That removed about eight percent of the dataset and cleaned up the variance noticeably.
Get the Full Details
What To Look For When You Start Digging
Most beginners look at setlists and count songs. That's not useful. Instead, focus on transitions. Phish is a band where the space between songs matters as much as the songs themselves. A show where Trey walks off stage between "You Enjoy Myself" and "My Mind's Got a Mind of Its Own" tells you something completely different about the evening than a show where they launch straight through. Another thing people miss: the second-set opener. This is where the band signals intent. If they open the second set with a deep cut from the '90s catalog, they're usually in exploration mode. If they open with "Run Like an Anchor" or "The Lark in the Morning," they're often settling into a groovier, more consistent framework. It's not a hard rule. But it's a signal worth tracking. I also started logging which songs appeared in which set position. "Manteca" in set one versus "Manteca" in set two produces different crowd energy curves. The band knows this. So do the tapers. So do the people in the front rows who somehow always know when the transition is coming even though they can't see the stage.
Advanced Patterns That Don't Make Sense On Paper
Here's something that doesn't appear in any book or interview: the relationship between temperature and setlist length. I know this sounds ridiculous. But I found a correlation between outdoor summer shows in venues without climate control and slightly longer average jam durations, especially in the second set. The band seems to lean into slower tempos when the room is hot. Maybe it's physical. Maybe it's psychological. I don't know. But the numbers don't lie. Another pattern: the "recovery set." Phish occasionally plays a show where everything feels slightly off—maybe a tuning issue, maybe a minor injury, maybe just a mental fog—and the setlist reflects it. These shows tend to have shorter jams, fewer segues, and more traditional song structures. They're also some of the most interesting to study because they reveal what the band considers essential versus what they treat as optional. You learn a lot about a band's core vocabulary when you watch them strip it down.
The Practical Workflow
If you want to build your own analysis, start simple. Pick a single tour—a winter tour is easiest because the setlists are well documented—and export the data. Use a tool like Google Sheets or Excel. Create columns for date, venue, song list, and any tags you care about. Then add a column for total runtime and one for average jam length. From there, you can sort by jam length, by song density, by venue type, or by any other variable you find interesting. One caveat: don't trust audio quality ratings too much. Phish.net and other sites tag audio sources with quality notes, but those tags are crowd-sourced and sometimes inconsistent. A "B" rating from 1993 might be a decent SBD mix, while a "B" from 1998 could be a damp cassette tape. I learned this the hard way after spending an afternoon analyzing what I thought was a significant drop in average jam length during March 1994. It turned out the entire month's data was skewed because three different tapers had submitted lower-quality recordings that had shorter runtimes due to tape warble. The fix was to exclude entries with ambiguous source tags until I could verify them manually.
Common Mistakes People Make
The biggest mistake is treating every show as equally significant. They aren't. Some Phish shows are one-offs with no real historical weight. Others represent turning points in the band's development. If you're trying to understand the arc of their career, focus on shows that connect to broader patterns—like the 1995 hiatus aftermath, the 1997 "Gamefinger" era, or the post-'04 reunion stretch. Those periods have clear musical narratives that you can trace through setlist changes. Another mistake is over-indexing on famous jams. Yes, the 11/12/95 "Divide and Conquer" is legendary. But if you spend all your time studying the ten most famous Phish jams, you're missing the actual story. The story is in the twenty-five shows between each legendary night where the band was still figuring things out. Those shows are where the evolution happens.
What This Approach Can't Tell You
Setlist analysis has hard limits. It can't capture the energy of a room. It can't tell you how many people were standing versus sitting. It can't measure the exact moment a jam shifted from exploration to resolution in a way that matters emotionally. For that, you need the audio. Or better yet, you need to have been there. Also, setlist data is incomplete for the earliest years. Shows from 1983 and earlier often have missing or guessed-at setlist information because the band didn't keep systematic records and the taper community was still forming. If you include those years in your analysis, expect significant gaps and be honest about the uncertainty. Don't pretend your conclusions about 1984 are as solid as your conclusions about 1996. Finally, there's the problem of subjective interpretation. Two people can look at the same setlist and draw completely different conclusions about what the band was attempting. That's not a flaw in the data. It's a feature of the music. Phish is a band that resists easy categorization, and any attempt to reduce it to spreadsheets and charts will always lose something in the translation.
