One of the things that limits your number of colors is the existence of Booster Drafts and Sealed Decks. These days MtG makes packs that are 16 literal cards but only 14 playable cards. Which means that in a Booster Draft you end up selecting 42 cards after seeing 252 cards in an 8 person pod or 207 cards in a 6 person pod. This in turn means that pretty much everyone can make a typical draft deck in Magic because you only need to successfully end up with 22-24 cards in 2 colors plus colorless artifacts. Even if the draft has really gone to hell and you had to switch colors in pack two you should still have enough cards in your colors that you don't have to play bad cards. In a Sealed pool, you don't get the luxury of seeing cards passed by opponents, but you do get to open twice as many packs, so you have 84 cards to whittle down into a deck. Even if there's an even split among the colors, you're still looking at no less than 16 cards per color, so there should always be a two color combination that you could make that would have at least 8 cards to cut when making a deck.
It's important to note however how important the conceits of Magic are specifically to having that format work out. If limited required you to make a 60 card deck that wanted 36 playable non-land cards in it, that would be a complete non-started. Few draft or sealed decks could put together 36 on-color cards for a 2 color deck. If we were playing Hearthstone or Force of Will and no land cards went into the main deck, we wouldn't have enough cards to form a playable deck. If we mandated that all decks be one color instead of two, a significant fraction of sealed pools would have no legal playable deck and you'd have to be a draft super star to get enough cards to field a deck in draft.
What does this have to do with the number of colors? The more colors there are, the more strain there is on the limited format. With 5 colors, a sealed pool has a minimum of 16 cards in its best and second best color. With 6 colors, the minimum is a still-playable 14. With 7 colors, the minimum is a borderline un-playable 12. With 8 colors, the minimum is a literally unplayable 10. Note that the inclusion of playable neutral cards increases the
average number of playable cards for a two color combination, but not the
minimum. The existence of multi-colored cards decreases the minimum and the average. For a Magic-like game, the maximum number of colors that the limited formats as we play them today could handle is 6. And that's assuming we have a pretty Magic-like game, most especially that there are resource cards that we let people put in their deck from outside the packs and that players can freely mix and match 2 color-factions. Games like Force of Will, Shadowfist, Vampire, or L5R do not have satisfying booster draft formats at all.
Another consideration is that you're going to have to make a playable identity for each 2-color combination. Obviously, they aren't all going to be equally good at every table, and they won't all be equally good in abstract. But they should all be playable and they should all have a shtick. Sometimes Magic fails on this - in Battle For Zendikar the Green commons were just
bad enough that all the Green/X color combinations had the identity of being "not very good" and in Kaladesh the Blue/Red combination was noted for being identityless and shit (despite having a truly excellent combo card, that people mostly played as a splash in Red/Green or Blue/Green decks). But it's obviously the goal you want to have.
For 5 colors, the number of 2 color combinations you are designing the set for is 10. For 6 colors, you are designing for 15 combinations. For 7 colors you are designing for 21 combinations (which is already kinda silly). And for 8 colors you are designing for a truly implausible 28 archetypes. To imagine how this works, let's imagine that your two-color archetypes are named after tribes, and for this example we'll do six colors with Purple Mountains, Red Moors, Orange Wastes, Yellow Plains, Green Forests, and Blue Swamps. This categorization uses a different one-letter code for each color and basic land type.
- BG Fairies
- BY Merfolk
- BO Lizardfolk
- BR Ghosts
- BP Goblins
- GY Elves
- GO Bandits
- GR Foulspawn
- GP Elementals
- YO Nomads
- YR Rats
- YP Dwarves
- OR Zombies
- OP Orcs
- RP Demons
That's a workable set of tribes, though a few of them are just one flavor or another of beastmen or people with funny hats. But imagine how this works in a Draft. First of all, your set has to have support at common and uncommon for all 15 tribes, which is an enormous amount of the print sheet. For example, Kaladesh has 101 Common cards, so if each tribe got just 2 common cards in each color that it appeared in, that would be 60 out of 101 cards dedicated just to the themed cards, leaving just over forty percent for unthemed cards. Trying to do the same with 7 colors would require 84 on-theme common cards, which is literally too large for a small set like Aether Revolt (70 commons) to even contain. Even with six colors, the average number of each commons opened for each tribe in an 8-person pod is about 2.4 - so with only 4 on-theme commons for each two-color combination, a majority of draft decks are going to be more than half off-theme filler or random good stuff.
Note also the literal impossibility of filling out your curve with on-theme stuff in a draft deck. You're gonna need 2 drops, 3 drops, and 4 drops as well as removal and combat tricks, and your "on theme" commons can only fill four of those slots (and you are quite unlikely to get enough 2-cost cards from the on-theme portion even if you have a themed bear from both colors to draft - chances are there are only 4 or 5 of those cards at the whole table. As the only Merfolk drafter at the table there won't be enough Tide Lancers and Bog Singers for you to get such that you won't need to draft some random Yellow Nomad or Blue Fairy - and the fact that this is going to happen means you are also unlikely to get every single opened Tide Lancer and Bog Singer.
Now you can alleviate this issue somewhat by making some cards that have synergies with multiple archetypes. A good example is the Maulfist Squad from Kaladesh.
In Kaladesh, Black/White wants as many dudes on the table as it can get, Black/Green wants to put +1/+1 counters on things, and Black/Red wants random artifacts coming into play. So a dude that has the choice of coming into play with a +1/+1 counter
or a little artifact dude is a pretty good dude for all three archetypes (in Kaladesh limited, Black/Blue is "bad"). But obviously you can only go so far with that sort of thing. If you print an Orange spell that both the Orcs and the Lizardfolk players want, it drops the average pull of that card for either player to a little over 1 if both Lizardfolk and Orcs are being drafted in the same pod. If there are three or more drafted archetypes hunting for a card, the average number each player gets is less than one and the impact is... limited.
Bottom line: the structure of drafting and the limitations of print sizes means that six colors is an upper limit of what can be supported in a set.
-Username17