Why are the tactical ratings so high?

How much do you believe that puzzles are inflated by?

I believe that to be relative to each player. My argument was rather about puzzles being consistantly inflated by x amount, so I was wondering what stops them from lowering everyones puzzle rating by some safe amount (like 100 points) to make it more accurate and not cause problems, (like puzzles having negative rating like you suggested)

I know this is an old thread, but this is where the googles brought me and the discussion above is relevant.

Would it help reduce ratings inflation if the rating increase for a correct attempt timed down to 1 instead of the current 5 point minimum?

Ratings are a skill to number ratio. This ratio tends to stay constant because you gain as many points as your opponent (in this case the puzzle) loses. It also tends to stay constant because for new players the system creates (or removes) points when moving a new player to their correct rating. So to change this ratio (to inflate or deflate it) you have to remove (or add) points from the population. For example chess.com could subtract a certain amount (like 200) from every player and puzzle.

But what you say could also work. For example if solving a puzzle gains 5 points for you, but the puzzle loses 6 points, then that's essentially the same thing i.e. you're removing points from the population (both players and puzzles are members of the population).

This has a ripple effect because, e.g. the system will essentially undervalue that puzzle you solved, so the next person to attempt it, if they solve it they'll gain fewer points than usual, and if they fail it they'll lose more points than usual... of course just 1 point difference isn't much, but over time this would build up.

I think it's because of two factors, first is that during a standard puzzle solving session people have ample time to calculate variations. For example a puzzle rated 2180 had an average time of around 40 seconds. In a real game if that position comes up then people would first have to "know" that there is a winning shot on top of calculating all the variations, people instead go for normal looking moves which are good enough. This means that in a real game that position could take someone a good 90 seconds or more to solve and most games on here are very quick time control games. In a 10 min game i doubt someone would spend 2 minutes on a tactic which "might" not be there.

For people playing long time controls where time isn't a factor, the main issue is that they don't know if any brilliant tactic exists or not and that alone drastically reduces the chances of finding it. 

I think these two factors are why the puzzle ratings are inflated and most people won't find such tactics in a real game

And actually (thinking about my last post) maybe that's how ratings got so high... i.e. people who don't know anything about the rating system messing with how it works.

If there really is a minimum number of points won (without making it equal to the minimum number of points lost for the puzzle), then that would inflate everyone's ratings over time.