I do, of course, play plenty of games I don't invent. I like to learn from them as well, to figure out what they did wrong so I don't make those mistakes in my games. A perfect case study for this sort of thing is "Sim City: The Card Game", which completely flopped several years ago. I discovered that the publisher was selling display boxes of the cards off at ludicrous discounts and therefore bought a couple boxes; I enjoy the game, but that doesn't mean it's not broken.
Very small key text sucks
A major problem with the game is the reliance on memorization of the cards. A new player will be swamped, and a seasoned player with unfamilliar cards will be confused. This is because the two most critical scoring effects in the game- Add and Complex- are not clearly labeled.
Add effects give a specific description of cards, and whoever plays such a card later in the game gets a bonus because this card was previously in the city. Complex bonuses are similar, but much more specific. The problem here is that the only way an "Add" card is marked is by having the "Add x to y within z blocks" text in really dinky print within its text box. It's trivial to overlook it. Some cards have effects for everything in the city, without a range constraint, so every play results in every text box on the board being visually scoured for bonuses. Although a much more realistic situation is simple frustration- becaus it's almost impossible to see the label unless you look explicitly for it.
This could easily have been fixed by equipping cards with large icons that can be easily spotted, informing players that the text box must be checked.
I am not colorblind. Actually, I'm more sensitive to color than most. That doesn't mean that I don't care that a game relies on color- I don't make strong mental associations to color, so I prefer to have a symbol as well. For that matter, there are plenty of people out there who are colorblind.
SimCity relies on the color of the pastel background of the stat box to say what phases the card can be played in. No problem once you get used to it, but one of my suitemates last year- Patrick- couldn't tell phase 2 from 3 from 4 because light pink, light green, and light tan are nearly identical to red-green colorblindness (his style). Green-yellow colorblindness would also have trouble between phases 2 and 3 (green and tan, respectively).
It's not the same issue, but those stat boxes are terribly placed for doing the extra duty of being a phase marker. They're at the bottom of the card; something like phase information really needs to be at the top.
Minimal player interaction
This is the big one. This is what really hurts a game. Players do not have any effective way of screwing each other over. The best way to play is to maximize your score with each move, because there's no way of knowing what your opponent might be able to do. Maybe xe'll be able to play on the huge sprawling Farm complex you just made ten points more valuable. Or maybe xe'll opt for the Residential zone- all because of the cards xe has in xir hand. There's no way to know, and even if there was, there's very little way to control.
High luck factor
There's very little control of the game in general. Optimizing placement so a zone will produce maximal points helps everybody, not just you. It's pure luck as to which cards you'll get, allowing you to take advantage of which bonuses- and the flow could go to someone else at any time from being shuffled a certain direction, and there's no way to know or predict. There is insufficient strategy in the game.
This applies only to a trading card game, like SimCity tried to be: it's not interesting to collect the cards. You do not open a booster pack and think "Wow! A Dairy Farm! Just what I needed for my deck!" That doesn't happen. The only exceptions may be Power Plants and Oil Refineries.
Everybody shares a deck. It's the only way the game works. But this means that more cards does not give you a competitive advantage; it lets you build larger and more intricate decks, but it's not like the Fireball you needed to complete a combo or the Solarion needed as the capstone for a deck. Once you have enough cards for a good game, there's nothing pressing you to get more; it doesn't have that crack-like addiction that Magic: The Gathering can inflict.
I guess the two things that hurt the game the most are the luck factor and the minimal player interaction, with complicated and illegible scoring a distant but still-severe third. It's interesting to observe that in games I like, there's very little luck and lots of ways to mess your opponents up. It's the nature of competition: the interplay between players, trying to advance your own position without letting anybody else get ahead.
I also like hidden information. It's fun to bluff. SimCity has that with closed hands, but it's not meaningful hidden information; there's a lot less you can do with a handful of random city blocks than with a handful of Magic: The Gathering cards, and the secret-until-it's-too-late moves of my newest (as-yet-unnamed) Piecepack game create wonderful uncertainty and ludicrous bluffing opportunities.
I actually had somewhere I was going with this, but I guess I sort of forgot it. I enjoy the SimCity card game, really- but it's good to learn from everything that contributed to it being a total and complete flop when I intend to make more games of my own...