Shooting the Moon is a tabletop storygame about a romantic triangle, which I first heard of on Sam Kabo Ashwell’s recommendation (and indeed played with Sam). The two suitor characters compete to get something (probably love and affection, but conceivably something else) from the third, “beloved” character.
The most standout aspect of the game is the character creation process.
This sounds like a slam, but isn’t: what I mean is that it does a great job of setting up characters with clearly defined conflicts and ways of interaction, giving the players lots of hooks to work with in putting together scenes. The beloved is created first, and given a bunch of positive traits. Then you write down near-synonyms and antonyms for three of the beloved’s traits, and these are assigned to suitors. So for instance, in our piratical game the Beloved had the traits “Honorable thief” (creating the related traits “Altruistic” and “Villainous”) and Literate (creating “Educated” and “Unschooled”), creating a villainous educated suitor and an altruistic one with no learning. As a next step, the players can add a modification to each of the suitor traits, so my character was refined to a villainous-but-loyal, educated-but-superstitious character. (At this point I made him a ship’s surgeon with a habit of making protective amulets.) Each character also gets a conflict, as well as a person, place, and thing that are touchstones and can be brought into the story at any time.
Most of these mechanics aren’t terribly different from elements found in other games — Fiasco playsets provide objects that are supposed to be important to characters, for instance, while many games assign characters some conflict to resolve — but Shooting the Moon excels at creating these characters as a set, with interrelated problems and concerns that are likely to feed off of one another.
Probably thanks to this, I found it easier than with most systems to improvise a somewhat coherent story — the narrative we played had a clear arc, and the characters interacted extensively. Contrast Shock and A Penny For My Thoughts, both of which are really interesting pieces, but tend to alternate focus between different characters without a lot of significant interaction between protagonists — instead, players typically wind up as antagonists or supporting roles for whichever protagonist is currently the focus.
Likewise, I thought the person/place/thing creation was just the right degree of useful. There was nothing in the mechanic that forces anyone to incorporate these items into particular scenes, and in practice we made heavy use of several while largely ignoring others. But they were useful when it came to providing improvisational support, when one’s mind is casting about for someone to bring on stage or an appropriate prop.
Some issues that arose with our session (but might not arise again another time, I don’t know):
— scene framing can be a bit of a challenge. This is really a game about interpersonal conflicts and connections, but once you get rolling on some kind of framing plot, it can be hard to avoid the urge to play out scenes that seem to be demanded by that plot. For instance, our story involved a major conflict in which the Beloved was carrying out a piratical revenge, so there was a temptation to play out battle scenes etc., rather than the scenes before and after battles that were actually important to us.
In practice, this meant that there was one rather improbable battle scene that characters mostly spent conversing, attempting to cast spells, and even singing a very loud ballad — because the combat, especially combat with nameless NPCs, was largely irrelevant both mechanically and thematically. On the whole, I found this a nice change of pace from games where you spend 50 minutes rolling exact combat effects and 2 minutes on role-playing some characterization during or at the end of it — but it takes some narrative self-discipline, possibly, to identify and move along to the right part of the story. This might have gone differently if we’d picked a different genre. As Sam put it: “When things are character-driven, you frequently end up in Kalamazoo at an all-night steak-waffle house.”
— conflict resolution is done in such a way that ties happen a lot: everyone rolls some number of dice, and whoever has the top score on any face wins. So if I roll 4 4 2 3, I lose to 6 1 1 1 despite having a larger total; if I roll 1 2 4 6 I tie with 6 1 1 1 ; and if I roll 6 6 2 4, I win beat 6 1 1 1. Given the system, I think it’s desirable for ties to happen sometimes, and probably more often than they would if a tie were defined as “players have the same total score”; but as ties lead to a somewhat cumbersome resolution in which characters can raise the stakes and then are pretty likely to lose anyway, having lots and lots of ties felt like it impeded the narrative pace. Someone who has played more would have to tell me whether our experience was anomalous, though.
Emily Short's Interactive Storytelling 
Our session was pretty anomalous as far as ties go. But play doesn’t suffer significantly if you just reroll ties, or instant-resolve them in some other manner. (People end up with more points, but who cares, really?)
Part of this: everybody seems to play this game slightly differently, and there have apparently been a number of fairly major revisions to the rules, which aren’t written in a comprehensive fashion in the first place.* An experienced player I spoke to had never heard of the Second Pool rules. The people I learned this from, also storygame veterans, thought that ties were broken by the next lowest die (6 6 5 5 2 beats 6 6 5 3 3). I kind of like the possibility of tie mechanics and saving losses with the Second Pool, but it’s definitely better if it doesn’t come up too often.
* For instance, the book never mentions at which point in character-creation you should decide who’s playing the Beloved, which seems like a pretty crucial piece of information.
And re: scene framing: this is another reason, I think, why being nailed down to a familiar genre is important. If you have the general shape and pace of a plot in the back of your head, things tend to slot into place without too much effort, and you can focus on character relationships; if your plot doesn’t have a well-defined shape, or its shape doesn’t quite fit into the space you have available, you have to spend more time struggling to get it pointing in the right direction, and less on flirting and fallings-out.
I think that the plot we ended up going for was a bit too big to fit into the space; we were going for something shaped for a doorstop-sized adventure novel, with side-treks and many stages, when the ideal size is something more like a short movie or a long (or dense) TV episode — too short, if anything, for nine-and-a-bit scenes, so that you have space for a scene or three that don’t do much plot work but advance character/relationship stuff.