There was a little bit of interesting discussion over on BoardGameNews on the topic of games with convoluted, boring, or anti-climactic scoring. It got me thinking about a few things.
I think of some kind of scoring or endgame reckoning as an inherent part of the boardgame form. Games are structured player interaction, and specifically measurable goals are a natural part of the structure. Still, could you design a boardgame without any scoring at all? I don't mean by this just not tracking points, I mean no game-specified objectives of any sort: no idea of player victory as in Pandemic or Lord of the Rings, no achievements as in Power Grid: The First Sparks, nothing. Just a set of rules that defines a system through which players interact and something that specifies how the game ends, and then the players discuss how it went.
The answer is clearly yes, but it would be hard. Republic of Rome is already somewhat close. I can imagine the Republic of Rome rulebook rephrased in such a way that the victory conditions simply became game-end conditions, with the players free to interpret them however they want. In actual play this already happens, as the game's victory conditions are usefully vague on a number of key points. Players are left to figure out on their own how much they value the Republic surviving but them not winning, as opposed to the Republic falling. Also, because the routes to victory are so elusive and opportunistic, players will spend a lot of time pursuing goals of their own choosing in the hopes that it will ultimately set them up in some way for a big play. Republic of Rome wouldn't quite work if you took away the entire concept of winning and losing, but it's conceptually close enough that I could imagine something that would.
This leads me to my main point: boardgames, just like all the other media we consume, are not about winning and losing. Yes, performance evaluation is part of the form; but it is no more or less fundamental to the success of a boardgame than it is to a roleplaying game, book, or movie. In roleplaying games there are no victory conditions, but in reality players pursue well-understood core activities which present characters with challenges they are attempting to overcome, where success is victory and failure is a setback, and that sense of "winning" or "losing" drives the tension of the game even when it is not explicitly said. That tension of anticipation or expectation is what is important. That we have goals, and that pursuit of those goals creates tension and drama – even if the expectation is failure (Fiasco). The goals needn't be explicitly specified, but it helps avoid inevitable misunderstandings when they are at least sketched out.
When thinking about opaque, implied, or non-existent scoring systems, ask yourself how it heightens the tension or drama of the game. In the case of 7 Wonders, it doesn't – we pretty much know the couple people who are going to be within the margin of error of winning well before we start the tedious aspect of counting points. Having so many categories to tabulate doesn't add drama because it doesn't interestingly affect game choices, and ultimately only a couple early-game factors make or break your chances: getting a good production engine set up that works in synergy with resources you can buy from your neighbor, having a reasonably steady flow of cash in trade, and not getting stuck in an arms race.
Agricola's scoring system, on the other hand, does heighten the drama of the rest of the game. If Agricola can be said to have a theme, it is of being spread thin, of having to build a farming empire when you have to spend much of your time just putting food on the table. By forcing you to diversify with a large number of potential penalties for failing to achieve minimal goals in a host of areas, Agricola's endgame spreadsheet does serve a purpose in driving your decisions in support of the game's theme. Also, because Agricola manages its tension well throughout the entire game by dramatically shortening the harvest cycle over time, scoring is rarely anti-climactic even when we have some idea of who has a chance to win and who doesn't, because in the end it's nice to know how we've done against the difficulties the game system has thrown at us in addition to how we've done against the other players.
There are similar arguments to be made about open scoring (like El Grande) vs. closed scoring (like Small World), or straightforward scoring (Through the Desert) vs. indirect scoring (Samurai). In each of these cases the choice of scoring dynamics serve the overall goal of maintaining game tension. El Grande's open scoring adds tension to its somewhat constrained on-board tactics, while open scoring would drastically reduce the game tension of Small World, where board play is much more open. In Samurai, the slightly opaque scoring system both supports the game's overall theme of indirection and heightens game tension by making theoretically trackable hidden information in practice very hard to track. Compare Samurai to Samurai: The Card Game, which is essentially the same game. Making scoring open in the card game heightens the tactical appeal, but slackens the game tension. We play the card game with screens.
The important thing here is that it's the underlying tensions and how they are managed over the course of the game that are crucial. Scoring simply as performance evaluation or identifying winners is not interesting. You have to actually care about the game for winning to be meaningful so if you're going to score, the way you do it has to validate the game experience. How it's done matters a lot. Open versus closed scoring, for example, is usually not just a matter of which way players prefer to play with a simple house-rule bringing things into line – one way will usually work much better in terms of maintaining the game's focus and drive, and will be the right answer.