OTB Ratings -- Where to go from here

First, two quick anecdotes, and then my question.

1.  I played chess from age 12 to 15, entered a few tourneys, and reached a USCF rating of 1335, and then left for other pursuits.  Like many, I got turned back on to chess after watching the Queen's Gambit, and, jumped in with both feet about six months ago.  This was after a 49 year layoff!  I've learned a tremendous amount over the past six months, and, last month, competed in the US Open (in Philly, along with 1000+ other competitors).  I was in the U1400 section (they dug up my rating from 1973!) and, as such, all of my competitors were in the 1100's.  In 9 rounds I scored 2 points (one on a forfeit, and one where I was seriously behind, but the other guy didn't notice his clock running out).  Essentially I lost all eight games that I played.

I emailed (noted chess instructor) Dan Heisman: what's going on?  I know so much more about chess than I did 49 years ago now.  Is it my age?  He said that, yes, brains work a lot fast in your teens, but he suspected that almost everyone there, especially the lower rated folks, are underrated, because they've been learning chess and playing on-line for 18-months (covid) and so their ratings are all from before they learned a lot.

And, indeed, these 1100 players know *so* much more than I did when I was a teen rated 1335.  I was never as good as these 1100 players are now.  (Just one example: a 9-year old played a decent KID against me.)

2.  I read an article in the most recent Chess Review by GM Jesse Kraai was at a recent OTB tourney.  He said "Was the level of play higher?  Absolutely.  I don't have a specific number yet, and it will certain take some time to measure how much the level of play has actually risen.  But here is my anecdotal evidence, confirmed in my conversations with other titled player: it used to be that a 2000 level player would drop pieces if you gave them enough opportunities.  That no longer seems to be the case.  And stronger tactics seems to be true down the line.  I have students in the chess.com 1500 range and they are dropping pieces far less frequently than they used to.  Puzzle rush and the competitive surge of the 2020 chess boom are having their effect."

=============

So, here's my question:  if everybody (and by everybody I mean those actually competing in OTB tourneys -- or restrict it to, say, under 2100 or so) is better, what does that do to ratings?  How does that shake out?  It's like: if a rising tide lifts all boats, then relative to each other, we're all sort of the same.  So, do we just live in an era where a rating of "X" is much better than what "X" used to be?

When I play OTB and I play agaisnt 2000+ who are older than 30 I feel like they're much weaker than the under 18 under 1800 players

"So, do we just live in an era where a rating of "X" is much better than what "X" used to be?"

Maybe.  

I thought there was rating inflation, but maybe at the lower levels it's rating deflation.

If literally everyone gets better (and equally so, e.g. 50 points) then yes, ratings are unaffected.

If only a sub group gets better, then it's more interesting. Ignoring what are essentially edge cases, the number of rating points in a system is constant. So (ideally) if we play, and if you win 10 points, then I lose 10 points. So if you don't play in tournaments while your skill raises 200 points, what happens (after many games) is you gain 200 points and everyone splits the cost of those 200 points.

This is interesting when it's a large group because the group itself has to "pay" as well. What I mean is, if everyone in a large group became 200 points better, and so everyone in the rating system has to "pay" 5 points, so do the members of the group! So in reality their ratings will only go up 195 points.

Another way to say this is even when the points are constant, the point-to-skill relationship of a rating system can change. The people who came up with the math realized this and there are several mechanisms in place that combat this. For example you might gain 10 points but let's say I lose 8. That means the rating system itself has created 2 points from nothing, and so that "cost" wont have to be paid by everyone else. The system usually generates (or destroys) rating points in this way for new players or players who have been inactive for a long time.

---

In a roundabout way, this also answers your last question. Yes the skill-to-rating relationship of a system can change over time. So it's possible that players rated 1300 today are better than players rated 1300 fifty years ago.

I don't know of any hard evidence, but from time to time someone like you comes along and gives anecdotal evidence to this effect i.e. "low" rated players know significantly more than they used to. I've never seen anecdotal evidence to the contrary either.

Well, that's not really what he's asking.

Typically kids are underrated because they improve quickly.

Here's another great example.  I played in the Atlantic Open last weekend.  On the board next me, the following position arose -- black to play:

White was a 10 year old, rated around 700.  Black was a 21 year old, rated around 1100.  

Black made the wrong move here, followed by the right move for white, and white won.

After the game, the 10 year old explained to the 21 year old why his move was wrong and what he should have played.

Bottom line: there's no way this 10 year old is playing anywhere near a 700 level -- she's clearly much stronger than that.  But in the books, this 1100 rated player lost to a 700 player.

==========

PS: OK, obviously, black takes a pawn.  But which one maintains the draw, and which one loses?  Just as important: after black takes the wrong pawn, what is white's correct response?

(Wasn't there a show called something like "Are you as smart as a 5th grader"?)

Taking the a-pawn draws, as when Black picks up the b-pawn, his King is out of the way of his running h-pawn, meaning he queens first. Taking the c-pawn loses, as described in the diagram below.

Phew! Not something I'd expect a 700 to see!

Nice position.

I wouldn't expect a 700 to know how rook pawns promoting attack each other, or to think about the difference between the two choices (for both players).

I wonder what the rating is when players start often solving this position.

Depends on the setting... OTB I'm U1300 and went 3/4 in my own OTB tournament. I have a feeling most of the kids there with enough time on their clock and incentive to think it out would solve this. I believe I would. (just so you know I'm on the younger side here )

Your ratings here mean you're way over 1300 OTB.

I wonder if a real 1300 could solve it though. I'm not sure.

I mean, it wouldn't surprise me either way, I just don't know.

OTB I was winning all of those games but it took a lot of effort even against a 700 and 900... I think they could if they concentrate and are given it as a puzzle but in a real OTB game they wouldn't have enough time (we were doing 45|15 and I nearly flagged one...)

Meh, maybe you just have to adjust your mindset.

In blitz a 1900 beats a 1300 before move 10 because the 1300 blunders a million queens.

I played a 1300 OTB not so long ago. I didn't win any material until like... move 30. OTB is a lot different. Players can stay solid for longer. When they're low rated you just have to be patient and wait... they always blunder eventually.

Daily ratings are super inflated though. I'm not surprised 1600 daily is 1300 OTB... in fact I probably would have guessed lower.

Yeah it's a bit counterintuitive but I'm getting used to it. Generally they can't convert an advantage seamlessly but are really solid. However once their position breaks down its all over.

Depends if you play daily randomly (me) or put a lot of thought into it.

So they were under the pressure like what I'm under when playing friends in rated rapid (just a parallel) where going 8/9 loses points.

Yeah, your 1500 daily rating makes me think you play it like blitz.

I knew a retired guy rated 1300-1400 OTB. His chess.com daily was over 1800.

But remember he was retired. He loved setting his games up on a real board, and analyzing at length. After a few days of this he'd make his move.

Yeah, get some derivative quotes instead.

So, then, next question:  (@llama47, I'm looking at you!)

In my OP, I noted that I played a couple of youngsters who were clearly way better than their ratings.

I was poking around, and a number of them are already 100-150 points higher than they were when I played them (which was in the beginning of July).

So, here's my question.  Suppose I lose to two players who are rated x when I played them, and then three months later, both of those players are x+150.  Does that, or will that, change my rating at all?  (In other words, is there a re-rating of sorts ever done based on the changing strength of past opponents?)

No, that would be chaos.