Tuesday, May 22, 2007

Together We Can Get This


My best friend thanks y'all for the compliments. It was a good tournament, and I actually have some other games that are interesting as well.

However, even that short game showed some interesting things. It showed how not to play, it also showed how to take someone out, by attacking your opponent's KB2 square. Sometimes descriptive is better than algebraic. However, he made the Ne4 play, not to checkmate the opponent, but to win the exchange. It was when the mate was hanging that he took advantage.

But, I have what I hope is an interesting question. It is for me. Using the USCF rating system, and the general principle that a 300 point ELO advantage, means that you will win almost all of the time. What are the defining characteristics of the following ELO levels?

100
400
700
1000
1300
1600
1900
2200
2500
2800

Not just titles, but what do you feel are the defining things, the things that they know. The reason they can beat the people below them, yet lose to the people above them?

I think this will help in my search. I also think this is going to be a good general question to ask others, and I will post answers as I get them in a perma-link in the sidebar.

More Later...

4 comments:

Temposchlucker said...

I reackon a factor 3 difference in speed for every 100 ratingpoints. So a difference of 200 points means you are 3x3=9 times as fast.
At a rating of 1800 you can play a 5 minute blitz game without serious deterioration of your game. With 1700 you can play 10 minute games, with 1600 you can play 25 minute games etc.. The difference in speed is not because you think faster, but it is the difference between knowing and inventing.

hisbestfriend said...

Tempo"S",

That is an interesting take. Did you get that from somewhere, or is it a fresh take?

Is there a range where that figure is in the zone? Just asking, cuz 100 to 2000 = 3^19 which is an awefully big number. Big enough that the 100 player is about a 1500 player by the time the game is over :-).

Temposchlucker said...

It is an idea of my own. Here is a post of mine about it.
http://temposchlucker.blogspot.com/2005/02/from-bungler-to-crack-in-02-seconds.html

I don't reackon with ratings below 1100. It is an estimation ofcourse, it can be 2.3x or 3.5x or so.
When you differ 700+ points, you can't statistically win (within a lifetime that is)
That is true for 1100 and 1800, but also for 2100 and 2800. A 2800 wins with the same ease of a 2100 player as a 1800 player wins from a 1100 player.

It is difficult to find the exact figures. How do you compare the following:
For an experienced user it takes 0.2 seconds to start his screen wiper when it rains. A novice without an operation manual and without technical knowledge needs half an hour to find out how it works. The figures change dramatically when the expert needs 0.4 seconds in stead of 0.2 seconds. So the right figure is "much".

likesforests said...

You might check out this link for a breakdown of strength by rating:

Google Groups article