It seems appropriate to analyze Grid and Iron Fist simultaneously as they share a great deal of design. They both feature the combo attack system, go out of their way to limit randomness, have the same type of character progression, and offer a limited set of items to discover.
The combo attack system is one of the elements I'm more impressed with. It has turned out to be an effective way to reward player skill and hide the lack of random damage rolls. In Grid, the combos allow for distinctive weapons to exist. In Iron Fist, combos are primarily changed by leveling up.
Ever since this post back in 2012 I've struggled with deciding how to implement randomness. This design uncertainty culminated with the release of Android <3 Kitty last year, which had no real randomness in the combat. Since then I've been riding the current and not bothering to break away from my own norm. Both Grid and Iron Fist offer very little randomness in the way combat goes down.
Grid is my first game to feature internal character progression that isn't just tied to equipment. @Star Wars attempted to have a similar kind of character development, though it was poorly implemented. In general I'm quite pleased with how the level-up skills work in both Grid and Iron Fist. It fits the more "arcadey" feel that both games tried to have.
One of the major successes in both Grid and Iron Fist was the implementation of Binding of Isaac-style consumables. Players are only allowed to carry one consumable. This has worked out quite well because it mitigates the too-good-to-use syndrome that players often encounter and also allows for tighter balance.
In a game with consumables, players are often tempted to hold onto them until they really need them. Sometimes, in easier games, that situation never actually arises. However, in games where it does, the player having a treasure trove of potions trivializes otherwise challenging content. Of course there's always a middle-ground to take, but I quite like the extremity of only one consumable.
Equipment is a strange topic in regards to both Grid and Iron Fist. When I was designing equipment for them, I wanted to create a small set of items where every item is notably different. I'd say I succeeded in my own goal there, though I'm not convinced it was to the betterment of the games.
All items are equally likely to show up in Grid and Iron Fist. There is no joy to be found in finding an item once you're reasonably familiar with the games as it will be something you've likely seen a hundred times before. To their credit, those items are generally pretty interesting, but there are so few and they have so little variance that I'm not actually sure it matters how interesting they are.
While I can't verify this, I suspect that most people that play Roguelikes do so for the replayability. The parent of replayability, in my mind, is depth. The most popular of the traditional Roguelikes (NetHack, DCSS, ADOM, etc) have considerable depth and thus replay value. This depth is created by interesting interactions, mountains of content, and uncertainty.
For all the positive things I can say about my own games, I think they have generally failed to provide interesting interactions, mountains of content, and uncertainty. In fact, looking back at it now, it appears as if I designed in such a way as to avoid those elements.
I'd like to change that moving forward and create more rewarding, in-depth experiences. This change isn't to spite Grid or Iron Fist or even Android <3 Kitty. Quite the opposite, really. Each of them has allowed me to learn about design and expand my coding skill. Now I'd like to take those self-improvements and apply them to a slightly grander project. My current plan is to have the 2015 7DRL serve as a springboard for that project. Until then,