Testing the Test

Today I stumbled across the website Elometer.net. Here's how they sum it up:

Dear chess friend,

we kindly invite you to the Elometer if you are interested in a quick measurement of your chess skills!

To obtain a psychometrically founded estimate of your playing strength based on a comparison sample of previous participants, all you have to do is to solve the following 76 chess problems.

...

In all of the following positions, you are playing White, and White is to move. Note, however, that the best move is not necessarily a brilliant one. Some of the following positions are rather simple, but some are very tough. Just try to choose the move you would play in a real game, and refrain from only trying to find a combination; this will yield the most precise estimate of your Elo rating.

The Elometer is part of a research project conducted by Birk Diedenhofen and Jochen Musch from the University of Duesseldorf, Institute of Experimental Psychology, Germany.

It took me a couple of hours to get through the test, but that was with stops due to supervising a 7 year-old. I would recommend doing it on a desktop computer instead of a tablet (as I did) or a phone, as that will minimize finger slips.

How'd I do? They estimated my rating at 2200 with a 95% confidence interval of about 2065 to 2330. Oops! Well, it isn't 100% confidence, is it? The format of the test also kept me from my usual errors of transposing moves and dropping random pieces, so I know where the 400 point difference is, alas.

Also of interest in a 48 position endgame test. All you have to do is state if White, with the move, can force a win in the given positions. I got 39 of 48 correct. Give them a try, and let me know how you did in the comments. The endgame tests include lots of basic stuff as well as more advanced material, such as the famous Cohn-Rubinstein ending from the St. Petersburg tournament of 1909, and Euwe's famous 1940 study of the pawn skeleton from the Ruy Lopez Exchange Variation. (Yes, I recognized both of those without having to look them up, although I did need to double check dates.)

I found the endgame studies interesting enough to enter them into a ChessBase database for further testing in the future. Aiming for no less than 48 out of 48.