The maximum value of flag weight is capped (originally at 500, now at 750). This is presumably to keep people from being able to cast hundreds of "inform mod" flags per day.
That makes sense.
However, I'd argue that it would make more sense to artificially limit the number of flags a user can have and keep the flag-weight uncapped. This will allow for the sorting of flags by weight to be more effective.
Before the flag weight went into effect, Jeff shared Stack Overflow's flag weight distribution in chat. At the time the maximum weight was 250, with only 73 people at that level. This represents just 0.02% of the Stack Overflow userbase. Even if you were to take flag weights of 150 - 250 inclusively, you would still only end up with 342 people.
Granted, people were unaware of flag weight then and might be more encouraged to flag now, but that's still a very small group. Now that the maximum weight has been raised to 500, the number of people at the limit is even less, at 27 people (0.008%).
From that perspective, increasing (or removing, in this case) the limit hardly seems to be beneficial, because it doesn't appear that the highest flag weight levels are particularly saturated. Should we assume that there's no harm in raising the limit as you said, it's still really is hard to see point in doing so since the current system is representative enough.
However, I'm not convinced that it wouldn't be harmful. Flag weight is a measure of reliability, but not necessarily one of importance. If you could get infinitely high flag weight by marking stuff on /review, it might completely bury flags about things like post vandalism under a sea of flags about non-answers. Said another way, a report from the reliable-yet-infrequent flagger about post vandalism shouldn't be ranked so much lower than reports from the reliable-and-frequent flagger about non-answers.
It's hard to say how the current system handles this, and I'm sure that in general there really aren't that many flags. Still, it feels to me that setting the cap at a reasonable level (which the current one seems to be set at) allows for a good mesh of flag reliability and diversity at the top of the queue, and taking away that cap comes across as a bad idea.
In part, we need something where a large number of bad votes will still actually put a dent in it; inappropriate flags are a net-negative.
However, we are investigating options for letting this grow above 500 (note that the cap on "how any flag-votes can I get" remains in place, though - you can't get infinity-infinity flags...).
This is in part tied into Flag weight 500: How can you tell if you're still flagging correctly?
We're looking at a few options here; at the moment the strongest contender is some kind of asymptotic scale, so getting to the current badge level (500) remains roughly the same, but growth above that slows down. You can still take pride in your exceptional flagging acumen though - and who knows, maybe some Au in the future (thinking off the top of my head there; not a guarantee).
Thoughts welcome...
