I own two versions of the word game Boggle, one from 1977 and the other from 2003. While continuing to whittle down my game collection to a more reasonable size (it once topped 1,000 games) and after consulting with my wife, I decided there was no need to own two copies of Boggle.
But which one to keep?
The 1977 edition has sentimental value and a sand timer. The 2003 edition (the “Game Folio” edition) has a zipper container, an electronic timer and a sleeker, quieter dice box.
Then it hit me… perhaps the most important question is this: Are the dice the same in both editions?
Each version had a single die with a “Z” on it, so I compared those:
1977: A-D-E-N-V-Z
2003: H-L-N-N-R-Z
Wow! Huge difference. How about the dice with “Qu” (again, one from each edition)?
1977: A-B-J-M-O-Qu
2003: H-I-M-N-U-Qu
Another massive difference!
But were the letters simply rearranged, or was the actual distribution changed?
Answering that question would require a lot of work. Pencils. Papers. Spreadsheets. A visor with a built-in headlamp.
So I did what any sane human would do. I turned to Google. And I found that someone else had already done the work.
The person or people at Bananagrammer.com put together a wonderfully deep analysis of the changes in the Boggle cubes. I love them for that.
The entire article is well worth reading, but here’s the tl;dr version:
There are more Ts and Hs (among others). There are fewer Bs, Cs, Gs and Ks (among others).
Based on the Bananagrammer tests, the new letter distribution results in an average 12% increase in available words per game of Boggle. The average score increases from 128 to 150. The average word length increases from 6.6 letters to 6.8 letters.
Armed with this knowledge, I still had a choice to make. Which copy to keep?
The answer was now obvious.
Everything essential from both versions fits into the 1977 box, so I’m keeping both! (But still saving space by getting rid of the Game Folio edition container.)
Thoughts from other Opinionated Gamers:
James Nathan: OK, I’ve got a 96 Boggle Deluxe that I either inherited from my grandparents or bought in the basement of a used bookstore in Syracuse (where I first saw the things that led me to discover BGG) or somewhere else I can’t remember. (Even if I didn’t buy this at that bookstore, I did buy a copy of Body Boggle there.)
Boggle Deluxe is a 5×5 grid (only recently did I discover that there’s a Super Big Boggle that is 6×6). My cubes are:
In comparison with the Bananagrammer link, I’m struck by my two cubes that have 4 Es on them, and that a single cube contains six letters that only appear once: B, J, K, Qu, X, and Z! With 9 extra cubes, this edition really limits those letters – in the “New” edition from Bananagrammers analysis, each of those are on different cubes.
While acknowledging that how often a letter appears on a face is not related to how often it could come up because of multiple-faces on one cube, I was struck by a few differences between my set and the “New” edition, as well as the general distribution of letters in English words. For instance, even with 9 extra cubes, my set has 1 less B and 1 less V. (The largest increase in my set is 2% more R faces (11 versus 5).
But the general distribution of letters in the language is where things get more wonky, as B occurs in around 2% of words, but is 0.67% of faces in my set – one face of one cube. While I suppose that gives it a 16.67% chance of appearing, it seems incongruous that Y, appearing in 1.77% of words, is on 2% of faces in my game, with three occurrences on three different cubes! W is similar, 1.28%, but 2% of faces with three on three.
Some of the intra-cube correlations are interesting too. For instance, if a cube has a D on it, there’s an 80% chance it has an N or an O. If it has an F, there’s a 100% change it has an R or an S. An M will def. have an E. (BTW, my hands now spell definitely as “def.” by default as it’s one of those that I’ve never been able to get the spelling down on.) Anyway, a P will def. have an I on it. My Y cubes alway have Is and Rs. Yet, for the “Classic” and “New” sets, even these correlations are different. For instance, while the “Classic” set a D cube has a 75% chance of having an N and a 50% chance of having an O, the “New” set has no D cube with an N or an O! The “Classic” Ys are only 33% likely to have an I or an R, and the “New” set is 33% I, 67% R.
There’s something going on with H also. It appears on 4 of my cubes, but on 2 of these, it appears twice. I’m not good at statistics, but H had the biggest likelihood of appearing more than once on a cube where it appeared at least once.
Looking at some of the relationships from both directions probably sheds some light on something too (but I don’t know what). That is, for cubes with an A, there’s only a 25% chance it also has an M, but for cubes with an M, there’s a 75% chance it has an A. If it has a T, there’s a 22% chance it has a W, but if it has a W, there’s a 67% chance it has a T.
I’ve certainly had a feeling over the years that something was… amiss with my vocabulary when I played Boggle, as it seemed like it fell into certain ruts and I found the same words over and over; I chalked it up to some sort of recency bias in my cognitive processes, but perhaps the phonology overlords at Hasbro have shorted some long prop bets on how long -RY words or those that contain a B and a K will last in common usage!
(FYI, the dictionary we use as our Honorable Boggle Judge we picked out based upon which one had the quantum particles I wanted to be able to get approved. Muon was the tough one.)
Recently, I thought I might quit working and try to go to grad school to study biochar, but perhaps I’ll go back for some sort of statistics and linguistics combo program where I study Boggle letter distributions. (Also, I just realized, my B-cube means I’ll never be able to spell kwijibo!)
Joe Huber: Completely unrelated to Erik’s Dilemma — but I have two Boggle-related thoughts.
First, Boggle is nearly the only game I play with my wife. Not a physical copy, but an iOS implementation; we’ve been playing an average of about half a game a day for many years now. We do play occasional other games together, but not nearly so often.
Second, the Children’s Museum in Boston used to have — maybe still does; it’s been years since I’ve been there — a recycling store, where they sold various items that local companies no longer had need for. And Hasbro is a (relatively) local company. So they have a number of game pieces, including — cubes from Boggle. While some of the other bits were useful for game design work, I picked up just five of the Boggle cubes — to put in my game room, and spell out GAMES. FWLIW, those five cubes are: AEEGMS, AFIRSY, AEGMNN, AAEEEE, and CCNSTW.

This photo, taken by Opinionated Gamer Melissa Rogerson, caused some confusion as we discussed this article. Not about the young girl (Melissa’s daughter), who was apparently terrorized by the idea of playing Boggle, but by the distribution of the letters on the dice! Is that tray even possible? What about the tray on the box cover? Melissa lives in Australia… so is this yet another edition of Boggle?
This is the kind of hard-hitting reporting that pays the bills and keeps the lights on for us here at the OG. Don’t forget to tip your waitress!
(Seriously, articles like this is why I love being a part of the team. Well done, Erik!)
Boggle was one of the games I played a LOT as a kid with my mom and siblings. I ended up with the family copy when we closed up her home. I like Super Boggle, but there is something about the simpler Classic Boggle that makes for a fun and casual game to play, when you’re just not in the mood or are splitting attention with other tasks.
I like that the link included both the 4 word requirement for Big Boggle in the calculations and then added 3 letter words for a more direct comparison. I note there are a number of forum posts on BGG about letter distributions for each of the versions, also worth a quick read.
Pingback: My Mind is Boggled – Erik Arneson