I am 1500 rated chess.com and 1850 rated on lichess yet it feels like my actual skill level is of a player around about the ~700 rating mark. I often hang pieces, I often blunder mate in 1 and I often miss the most rudimentary of tactics. This causes me to lose 60% of my games so even though I'm being matched against equally rated opponents, there's a severe disconnect between the strength the website ranks me at, as my actual real life skill level.
When are chess websites going to start rating me appropriately? I am not a 1500 rated player, I am a 1200 rated player, maybe 1300 when I'm playing my best. I'm thinking to just sandbag my elo on purpose because the rating system chess websites use just isn't legit.
For context, just played a rapid game with this position Black to play.
That position and the tactic there is definitely NOT 800 level. That is for sure a 1700+ tactic. Let me explain why it's a 1700+ tactic
If you go deep in that position then u will realize why im saying this. Rxh4 is what everyone sees because the pawn can't take because of Qxh2 mate. However white does not NEED to take the rook. What if white plays Bd6 which attacks the queen and forces it off from that diagonal? Keep in mind that if the queen moves from the diagonal then white can take that rook. So what do you do now as black? if you move the queen then gxh4 and black is not "winning" anymore. One move which people might see is Rh1+ after Bd6 because if Kxh1 then Qh5+ followed by mate. However, again white doesn't NEED to take the rook, white can play Kg2 and now both your rook and your queen is attacked. Now there are only two "winning" moves, Rh2+ which forces the king back so that u can move your queen to h5(getting out of attack by the bishop on d6). Either that or you must see Qf5 which leads to checkmate in 7 moves(not so obvious). Keep in mind you would have to see all of this before playing Rxh4
Do you see what i mean? maybe i am overevaluating the tactic but i am very very sure that it's atleast higher than a 1700 rated tactic. You basically have to evaluate this variation in your head
Rxh4 Bd6(or Bd4) Rh1+ Kg2 Rh2+ Kg1 Qh5 and this will lead to mate. This is just the "winning" continuation. People might also give up the rook after Bd6 (say Qh5 after Bd6) and might think that the king is wide open with knight and queen near it so it will lead to mate, however a sneaky bishop move will defend everything. I had to look very deep into this so as to find the absolute winning move.
Not to mention there are other good looking moves like Nxf2 and something(Kxf2 is basically forced after Nxf2)
So basically before playing Rxh4 you would have to see Bd6 attacking the diagonal and forcing the queen off... then u would absolutely have to see that you can play Rh1+ because if Kxh1 then Qh5 after that leads to a mate in 2. However you would have to also know how to respond after Kg2 which simultaneously attacks both your rook and your queen and you can't even play Qh5 due to simple Rxh1. ALL of this is very risky since you are "hanging" material in every single line except the winning continuation.
However, I found Rxh4 in a few seconds. I’m way lower than 1700, and once I found the line, I only took a few seconds to calculate it. As you said, could you be overvaluing the tactic, but more than you think?
Two points:
1) After a bunch of games, most people should expect ~50% of their games to be wins
2) Rating is a function of your opponents and scaling factor does not matter. An easy way to think of this is if you just multiply everyone's rating by ten. Their relative position does not chance, nor does the % difference between them. What is important is what the number represents, which is the relative position of you vs. other chess.com users.