In my podcast I've said a few times how little I like custom dice.
I have to be honest, though: I'm starting to feel less negatively towards them. In due course, I might even think they're OK. But, generally, although they do enable certain mechanics and game options that are definitely pretty cool, overall I'm still opposed.
The main reason I don't like them is because they are an enabler of the sunk cost fallacy in which people continue to invest in a game or system, not because they enjoy it but because they've sunk too much money into it to back out now. Obviously, miniatures and stacks of expensive, full-colour, edition-limited rulebooks are the main manifestation of sunk costs in miniatures wargames. But dice are in there, too.
It's funny, in fact, that Games Workshop - arch exploiter of their customers' sunk costs - has resisted a shift to a custom dice mechanic as far as I can tell. I mean, you could change up the distribution of probabilities on the dice faces for each new edition, forcing customers to buy entirely new dice every time they buy a new rulebook... I know: I shouldn't give them ideas, should I?
The dice in the image above are from Star Breach, Star Saga and Aristeia! which are all good, solid games. The Star Breach dice, at least, also double as d6s if you look closely enough, which is commendable. Otherwise, they are a lot like Bolt Action orders dice, which I don't count as custom dice - they're really just six-sided counters, not intended for rolling. The Star Saga dice, meanwhile, each have four* symbols on them in the ratio 1:2:2:1, which is mathematically pretty interesting and hardly difficult to replicate just using regular d6s (1, 2-3, 4-5, 6). The Aristeia! dice, meanwhile, also have three different symbols, in different proportions, depending on the type of dice: 0 (1), XY (1), X (1), Y (1), XXZ (1) and XYZ (1), where X, Y and Z are different symbols (and 0 is no symbol).
The Aristeia! dice are both the cleverest and least likeable of the three examples, but that doesn't mean I don't like them. The symbols are clear and comprehensible (blast = offence; shield = defence; exclamation = special), and the ratios are cleverly distributed in such a way that you could re-purpose the dice into an entirely new game (as Corvus Belli did with their Kickstarter-exclusive Infinity: Defiance sci-fi dungeon crawler). Buuuut, they still aren't games that a regular player could easily re-purpose to some pre-existing game, so they still shore up the sunk cost fallacy.
The acme of this complaint are the Fantasy Flight Star Wars RPG dice:
Look at these monsters! And now let me tell you that (1) some symbols cancel other symbols; (2) you typically roll more than one dice of several different shapes at one time; (3) some symbols require the GM to make up a unique story twist on the spot, (4) there are eight different symbols on these dice, and (5) FFG went and re-designed them for their generic system, Genesys:
This is practically a crime against humanity in my book.
It has to be said, though, that some people love them.
In fact, some people find symbols easier to process quickly and intuitively than numbers, and I'm certain that there's some interesting research to be done, if it hasn't already, into this subject. People with, for example, dyscalculia might well find abstract symbols easier to process than numbers. And the matching/canceling mechanic is hardly one I'm in a position to criticize given that Horizon Wars games use it in opposed rolls.
Ultimately, to me, it comes down to a question of economics and sustainability. A numbered dice will always be useful, one way or another, in almost any quantity. But a custom symbol dice, eventually, will become useless when the game itself ceases to be played and will become a piece of waste. Or, they form part of the marketing sleight of hand that uses behavioural economics to keep customers buying products to sustain an otherwise unsustainable system. And, to me, those are both big negatives for custom dice.
*One symbol is the lack of a symbol - a blank face.