The “Grid Method”
Where we found it: There’s an entry for it on Invisible Castle, the indispensable dice rolling website. While the site gives credit to a user on a forum, the links to the original forum no longer work.
What is it: A way of generating ability scores for a new character. Rather than using a standard array, simply rolling dice, or a point buy system. The grid method has players generating a 3×3 grid of numbers (by rolling 4d6 nine times) then labeling the rows and columns with the six ability scores. Players choose one number from the corresponding row/column for each ability score, and each number can only be used once.
Why is it important: This method is a lot of fun, and lets players make interesting choices about their character right from the ability score phase. Rather than simply applying a min/max formula to assigning ability scores, the grid method leads to trade-offs and often times surprising results, which can create characters different than just the perfect mathematical representation of a class or archetype. There are strong puzzle and discovery elements to crafting a character this way.
We love the method, so much so, that some within our play group might never use another means of character generation. Yet, in practice we noticed a few glaring flaws. First is that by rolling to generate the grid, there can be wild swings in the number set, making the choices either academic or potentially impossible. If you generate nine numbers, none of which end up being less than 12, you’ll probably create a vastly overpowered character (and that’s before you consider racial/class bonuses). Of course, the reverse is also true and you could end up with a vastly underpowered character.
Another issue is that because of how the ability scores are grouped (STR, DEX, CON on one side, WIS, INT, CHA on the other), the grid essentially splits characters into physical and mental types and is biased toward classes that embrace one of those types. So, fighters and wizards are easy to make, but classes that relies on strong scores in both categories – like a cleric or monk – are harder to create.
So, after more calculations than were probably necessary, and lots of tweaking, we present the Random Grid Method.
The Random Grid Method
The changes that we made were aimed at solving the two glaring flaws we mentioned above. For the numbers, we decided that rather than generating numbers, the grids should all use the same number set, but that the random arrangement of the numbers was paramount. As for the numbers themselves, we decided on using the standard array as a base, but since we need 9 numbers for the grid and not six, we tweaked from there.
A standard array for D&D Next looks like this : 15, 14, 13, 12, 10, 8
Our expanded array for D&D Next looks like this: 15, 14, 13, 12, 11, 10, 9, 8, 8. *
As you can see, we’ve added three new numbers, an 8, a 9 and an 11. This came from testing out number sets that would maintain the average of the standard array, while providing a bit more “garbage” that would force players to make real choices in their grid and not simply always get to circle 15-10.
*Note: You can use this method for 4th edition as well, just use this expanded array: 16, 14, 14, 12, 12, 11, 10, 9, 8.
To assign numbers randomly simply give each spot in the grid a designation 1-9 (left to right and top to bottom) and roll a d10 (ignore 0) for each number to place it in the grid.
Once the number issue was solved we moved onto the problem of the ability score bias. The solution is to just make their placement as random as the numbers in the grid. The simplest way to do it is to designate each ability score slot 1-6 and then again, roll a d6 for each ability score’s placement.
When you combine randomly assigning the expanded array with randomly assigning the ability score labels, you end up with grids that lack a physical/mental character bias and that will generate characters with ability scores within the expected range. Here are a few examples:
As you can see, sometimes the random assignment of numbers and ability scores will offer some interesting choices. If you were making a fighter from the grid on the left, STR will compete with DEX and CON for a score.
Use this handy PDF for step-by-step instructions on using the Random Grid Method. We also provide some sample grids to use – in case you don’t want to create your own – and a more detailed FAQ.
In case you were wondering, there are over 720 possible permutations of ability score labels, and there are 362,880 possible ways to distribute the numbers. Suffice to say, you probably won’t run out of grids anytime soon.