Borellus Connection Review

My in-person Baltimore group recently completed our campaign in The Borellus Connection for Fall of Delta Green. Or rather we mostly completed it as we skipped Saigon because I found it too intimidating to run and skipped Marseilles because I was moving back to Los Angeles and we needed to wrap it up. But even having skipped two chapters, it still took about 18 sessions of two hours each to complete.

My overall review is that it’s a great campaign with strong world-building and characters that tightly integrates a ton of Lovecraft and Delta Green lore but it’s not GM-friendly, especially by the standards of the OSR.

The good part is the story and the setting. It has everything you’d want for a 1968 Delta Green adventure. One advantage of a published Fall of Delta Green campaign vs a regular Delta Green campaign is that setting it 53 years before its publication date means that there were a ton of colorful historical figures who were dead and could thus be safely be integrated into the fiction. Seriously, like half the most memorable NPCs are real people, which shouldn’t surprise any regular listener of Ken and Robin Talk About Stuff. There’s also the related black humor of repeatedly being shown things like the construction of Olympics facilities in Munich that the NPCs are all really excited about but which you, the GM and players, know will end in tears.

But we also enjoyed lots throughout and unlike my Discord group’s Hot Springs Island campaign, I would call my experience with Borellus Connection a successful campaign and recommend it.

The many good things about this campaign are what make it frustrating that it was hard to run. Part of this is that when I’m not running Gumshoe I am running OSR which has lately emphasized information design so good that a GM can run a module cold. I had to read each Borellus chapters two or three times and I still had a harder time GMing it than I did running OSR module by Kelsey Dionne or Brad Kerr that I either skimmed for 15 minutes before the session or hadn’t read in months.

I remain unsure how much of this difficulty is that it’s intrinsically impossible to write an investigative horror campaign that’s as GM-friendly as a dungeon crawl and how much is that it could have been done better but the information design was suboptimal. I am confident though that it’s not just one and not just the other.

Let’s start with the information design. Borellus Connection definitely has good layout in the sense of just enough “graphic design is my passion” artpunk but not so much as to make it unusable like Mork Borg. Moreover, Borellus Connection does follow some basic information design principles and isn’t 1980s style wall of text. Each scene gives an introduction (which is effectively box text), then a bunch of triggers the players can activate to get clues and other resources. That’s great and similar to best practices in OSR information design.

But there are also readily fixable problems with the editing and information design. Some examples:

• On page 251-252, where does the “Radio Liberty” scene end and where does the “Talk Over Coffee” scene begin? The last two lines on page 251 are part of “Talk Over Coffee” but are placed before the heading, at the end of “Radio Liberty.” This is an unforced error as there’s enough room for these lines at the bottom of the first column on page 252.
• Page 227 says “see the sidebar.” There is a sidebar on page 227, but this text refers to a different sidebar, which is on page 228, which means it’s not only a different page but a different spread. Gavin Norman would have his freelancers fed to a gloam if they did something like that but it’s almost more common than not in Borellus Connection. For instance, the custodes (zombie) flavor text is on page 23 and its stat block is on page 24. It is a pretty basic layout principle that if you can’t fit something on one page, put it on one spread, not one leaf.
• Page 259 spills onto page 260. Again, there was plenty of space on page 259.
• Why is the connection from “Salafists” to “New Ways to Shout” presented as a lead-in rather than a lead-out. What is more natural for flow at the table? “You can get here from the other place” or “if you do this here you can get there?”
• Page 362 “some form of some form of glossolalia.”
• There needs to be a lot more cross-referencing. For instance, on page 397 it mentions the staff of Ógafracoth. Where the fuck is that in the book? It is not in the index, Ctrl-F of the PDF for “staff of Ógafracoth” doesn’t help, and it’s not in Ógafracoth’s stat block on page 38. (It’s actually mentioned on page 318, but good luck finding that in real time).
• Suppose there’s an investigative skill called Widgetry. Sometimes it says “Widgetry core clue, the doodad smells like a type of cologne only sold at the duty free shop at Heathrow” or “for a one point spend in Widgetry, the doodad seems a bit hinky” both of which make sense to a Gumshoe GM. But most of the time it just says “Widgetry, the doodad was salvaged from Karotechia stationary.” Does that take a spend or not? I would assume it would take a spend if not for the other places where it explicitly says “one point spend.”

All this stuff is what’s important about layout and editing for making a product usable at the table, even if what you notice on a flip through is the Photoshop filters on public domain art. There’s a lot of room for editing and layout to improve the usability of the campaign, but I think there’s a lot about the nature of the campaign that means even improved information design would only get you so far as investigative horror may simply be a harder scenario structure to GM than dungeon crawls.

And even a dungeon crawl can be hard to run. I’ve never run Halls of Arden Vul, but in thinking about how Borellus was hard to GM, I keep thinking about how Arden Vul is a legendary megadungeon, but not one that’s usually described as GM-friendly. There are information design improvements you can make, hence Joseph Browning’s bullet point notes, but there’s also just hard limits on GM friendliness contingent on trying to accomplish certain gameplay goals. Hence Beau Rancourt’s recent blog series on GM-ing Arden Vul had plenty of criticisms about how each room key is formatted but his post mortem concludes with this note:

Are you familiar with the phrase “good, cheap, fast; pick 2”? I think there’s a similar thing happening with mega-dungeons. “Interconnected, Dynamic, Huge; pick 2” feels right to me, but Arden Vul tries to do all 3.

Borellus Connection shows that the only think Rancourt got wrong was to use the word “mega-dungeons” instead of a more generic term like “published campaigns” because interconnected, dynamic, and huge is a big part of why Borellus Connection would still be challenging to run even if it had better editing and information design. My complaints about formatting and editing notwithstanding, the scenes are mostly easy to run when they can be treated as discrete units. Just like Rancourt’s interconnected, dynamic, huge lament about Arden Vul, the difficulty in Borellus Connection is mostly in when you need to connect a scenes to events and especially NPCs described elsewhere. Borellus Connection is merciful in that it mostly lets you focus on one mission (usually, a city) at a time, but the missions are also fucking huge. The longest chapter, Saigon is 66 pages long. While I noped out on running Saigon for my Baltimore game, I later ran it for my Discord game and it was as challenging as expected. The easiest chapter to run is Laos, which is short and linear, and the flight to LAX, which is long but linear.

I have found that Arc Dream Delta Green adventures are easier to run than Borellus Connection, even with the added cognitive load of converting from d100/BRP to Gumshoe. Part of this is that Arc Dream has better layout and editing than Pelgrane. But just as important is that I have only run Arc Dream one-shots. “A Victim of the Art” is 16 pages and “Operation Fulminate” is 12 pages. That’s really easy to run as it’s a small enough amount of material that you can keep it all in mind. Maybe I’d be overwhelmed and frustrated if I tried to run Impossible Landscapes, but I don’t plan to. And Arc Dream’s style has its own fail state, with “Kali Ghati” being way too much of a railroad.

The only thing I found that really helped was to give a chapter of Borellus to an LLM and then ask it questions when I got stuck. I mostly ran the sessions normally, but then asked the LLM’s opinion when I got stumped. This worked way better than anything else I tried (reading the chapters two or three times, asking an LLM to make a bullet point layout cheat sheet for each scene, using R/igraph to generate a viz of each chapter). I’m glad to have this solution as the substance of the campaign is great and my group really enjoyed it, I just wish it wasn’t necessary to have a baby god co-GM a campaign about a wizard seeking to become a god.