I was brainstorming about finding an optimal way to have a two-card system, in this case optimizing it by the number of pre-made images you have to make. Here’s what I’ve found, I’m not sure if I’ve done everything correctly but I’m open to anyone who wants to correct my math. As a forewarning, this system does use two images for each pair of cards as I’ll describe it, but I think with some creativity, it could be simplified to a one-image system.
I read about systems which split the information of two cards into their suits and values, instead of thinking of them as two different cards. So to start, I decided to have 16 images for suits: one for each of the possible combinations you can get with two cards. (so hearts-spades and spades-hearts would be two different images)
For the values, I did something a little different. Instead of having 13^2 = 196 images for each possible combination of cards (although you could totally do this), I wanted to make it even easier. Don’t even worry about the order of the cards, just have an image for each possible combination of two ranks (e.g. one image for Ace/Ace, one for 5/7, but the image for Jack/8 would be the same as 8/Jack because order doesn’t matter). We can encode the order by deciding whether we place the suit image before or after the value image in the memory palace. By doing this, we reduce the number of images to 13+12+…+1 = 91 images for card ranks.
Suppose we have the 2H followed by 3S. Then we’d use the images for HS and 2-3, but we’d put the HS image first because the lower-ranked card comes first.
Conversely, if we had 3H followed by 2S, then the images would also be HS and 2-3, but we’d put 2-3 first to indicate that the higher-ranked card came first this time.
Note: This also works for two cards of the same rank, you can put the images in either order because the suit images have order taken into account.
I do recognize that this isn’t a true 2-card system because we’re still using two images to apply to two cards, so I’ve essentially increased the number of images from 52 to 107 for a 1-card system at this point. However, I think with some creativity we can come up with 16 “attributes” to apply to the 91 rank-images, thus encoding everything into one image for a pair of cards.
For example, suppose we have some attributes like jumping, kicking, etc. If the object itself is kicking/jumping, that would mean lower-ranked card first, but it is being kicked/jumped on then the higher-ranked card comes first. Or we could do colors: if the object is red/blue, then lower-ranked first, but if it is bleeding red/blue onto its surroundings (or whatever other interpretation you can come up with), then the higher-ranked card comes first.
Please let me know if you have any thoughts or contributions as to what type of images/attributes would be easiest to remember, or which ones would work best with each other! I think this will be a nice alternative for people like me who want to try out a 2-card system, but don’t want to spend the time to make a “real” one.