On-the-fly Game Design (Why Don’t I Do This More Often?)

Hackin’ on Bananagrams

At a small game night last week, somebody brought out Bananagrams (not a typical game for a Board Game Night, but it was a nice, quick filler while four of us wrapped up a game of San Juan). We started throwing around ideas for modified Bananagram rules after a game or two and decided to try a few of them out. The game lent itself very well to rapid, on-the-fly game design since each playthrough took less than 10 minutes, and after a few iterations we were able to arrive at a ruleset that I think we were all pretty excited about.

It felt great to scratch my game design itch in a quick, loose, and collaborative setting like that, and we made some explicit plans to do more on-the-fly game design in the future (optimizing for iteration by picking simple, quick games). I’m honestly a little surprised that none of my game nights in the past have turned into game design sessions like this, since it seems like a logical next step (and was incredibly fun), and I can’t wait for more like this in the near future.

Rule Modifications

When someone says “peel,” every non-peeler quickly chooses one tile from their own area (maybe an X or Q that has you stumped), gives that tile to the peeler, then draws a new replacement tile from the center. So with 5 players, the peeler receives 4 new tiles (one from each other player) and everybody else has swapped an unwanted tile for a random new one.

At the end of the game, whoever has the most placed tiles wins (which might not be the last person to peel).

Trashing tiles (which normally increases your number of tiles) is not allowed. You just have to wait for someone else to peel.


This variant addresses two of the big problems I see with standard Bananagrams: breakaway leaders (addressed through rubber banding) & randomness.

Rubber Banding

You win by having the most placed tiles, and the only way to add tiles to your board is to peel. However, this altered peel, which I’ll call the “junk peel,” is a rubber-banding mechanic of sorts. It actually brings the non-peelers closer to being able to peel by letting them trade out a tile they don’t want (without increasing their number of unassigned tiles), while slowing down the peeler with a bunch of new letters all at once. Standard Bananagrams often results in one or two power players going on long peel-streaks, whereas this variant sees a lot more variation in who’s calling peel (the added inclusivity’s a nice perk for a casual/family game like this, and it also means that everybody tends to get a few chances to trade out unwanted tiles).

I was worried that the rubber-banding might be too strong and every time you said peel you’d end up with 4 X’s, Q’s, and Z’s, but in practice the tiles that get junked are more likely to be perfectly good letters that happened to throw someone’s vowel/consonant ratio out of whack. Even if the peeler does get some lousy letters, they can always trade them out when others peel. And eventually those X’s, Q’s and Z’s will find a home on someone’s board and leave circulation, so it won’t devolve into playing hot-potato with lousy letters.

Note: One of my favorite episodes of the boardgame design podcast Ludology (ep. 3: catch-the-leader mechanics) touches on this topic.

Downplaying Randomness

Standard Bananagrams has an issue with randomness. A streak of all consonants or a cluster of tricky letters can be a death sentence. You can get rid of lousy letters, but at a significant cost (netting 2 additional tiles). However, in this version, while the game still features lots of randomness, the junk peel mechanic minimizes the impact that a bad draw or two can have.