Estimating My Chess Rating with

Try it yourself:

Apologies for the sporadic audio issues… I’m still learning!

Social Media:


  1. I got 2139 elo, more or less in line with my classical rating. My blitz rating is crap though, so if this were a timed test I would've gotten 1800 elo tops.

  2. I currently only play blitz games and my rating is about 1000. Got about 1700 elo for the test result, feels different solving puzzles compared to versing a real player.

  3. 8:34 "This is mate in 2" Sees it in 1 second. I had to rewind 3 times to see it.

  4. 80% of comments: Woah it says I should be (insert rating 300 points higher than reality)
    20% of comments: They (are/aren't) using the ratings you put in.

  5. I got 1583 with 95% confidence interval. I cant compare it to anything official, but it is close and a little higher to my lichess rating. It was good puzzle practice regardless

  6. I did this test and it gave me an approx. rating of 2210 when my online accounts are rated about 1950. Interesting how i managed to solve most of these problems (very proud) but i did miss a few easy tactics which is disappointing.
    Edit: Btw the problem at 53:23 this one is the only problem where im 90% sure that bf1 is not the winning move but instead bg2!

  7. Mine was way off – it estimated me at 1911, when I'm probably 1400-ish player on my best days.

  8. First time visiting site: already has an estimated elo

  9. Thanks for posting this Eric, didn't know the site and me and my nephews love it!

  10. Hey great stuff 😉 keep doing your youtube videos I like them ^^

    30:27 just a little question. Why not king c3 ? Black King cant run fast enough and he cant push the pawn. So if we walk the long way or the short way with c3 is pointless, isnt it ?

  11. My USCF rating is around 2000. This test estimated my ELO to be 2230. I got all the same problems you got but in a different order. I didn't give them any data on my ratings or games played and such, as I was suspicious they'd use that to tweak the algorithm's estimate.

  12. First I was going to comment that we had 30 different answers or that I had a 1800 something rating, but then again, WTF YOU'RE 24????????????

  13. It says there is a 100 percent chance and it is a 100 percent guaranteed that I am at least 3000. It didn't give me a upper limit though.

  14. 2266 – way above my actual rating, which is around 2000 xD

  15. I got 2209, while my USCF rating is 1957. This might be because I have lots of tactics experience, or maybe I'm just underrated. I remembered that problem 26 was something you and Kostya pointed out in a recent video, I wouldn't have gotten the puzzle so quickly without the video! Also on puzzle 32, I noticed that you calculate far deeper than I did. I played the same Rh8 but I didn't even consider the difference between Qh1 and Qh2. Thanks for making this video! You should do tactics trainer sometime, it seems like your audience is anywhere from 1200-2000 and would appreciate hearing more about your thought process.

  16. Got a 1883. Havent played any tournaments but I have played online and my rating is like 1100 or something on blitz haha

  17. 2174 with 95% confidence.
    Did it on 2AM which might took me couple of rating points but i got much more than my actual rating (if you assumed they use your rating my national is around 1800).
    I agree with your statement that if you don't play the whole variation it makes the elo analysis wrong.
    Basically i think that tactic trainer and alike are helpful for chess improvement but as long as you don't know how to get these thematic position onboard it will not estimate your rating correctly.
    Specifically in this one they didn't have many defending themes nor complex tactical themes like rook lifting ideas or positional piece sacrifice especially rook trade with knight or bishop (except one that was pure tactically) so many complex ideas didn't get in. That makes me believe that if lets say Magnus Carlsen would do that he would get 2500 ELO at most which makes it bit irrelevant.

  18. "I'm gonna go quick, I don't want this to take all day…".

    An hour and a half later….

  19. This reckons I'm an 1800. Like….. no. I'm a 1200 at best.

  20. I'm 1690 provisional at the moment. I got ~2200 on this test, and assumed it was just a failure of a test since I'm nowhere near that strong. Interesting to see this video and you saw the checkmate patterns so quickly 🙂 That's why you're an IM and I'm not, I guess 😛

  21. Props for getting through this entire thing. I got about halfway through and said "forget this!". 76 positions are tough to get through in one sitting.

  22. Elometer: tell us your rating and we will estimate your rating!

  23. I don't play chess against anyone except my friend but I play different computers alot as well as alot of puzzles. I play GO the same way and spend more time playing it than chess. I love puzzles!
    1877 with 95%, This is with a miss-click moving a rook to an illegal spot.

  24. Thank you Eric for lecturing while solving questions and helping us mediocre players get better. (I took the test and got an 1800, 200 above my real elo)

  25. The first page was just an exemple, not a distribution

  26. Got a much higher rating than UCSF. Several interesting elements: Only the first move is checked for accuracy so faulty calculations can be covered up by good intuition. I have done puzzles on Lichess with similar themes even if they were not exactly identical. OTB pressures of time and otherwise are a very different ballgame for many of us.

  27. my french national rating is 1240 but I got an estimate 2017, not that much accurate x)

  28. Thanks for the great video! I did the test too, and afterwards I found a website with the solutions. So a couple remarks:

    55) The solution given is h4. The idea is probably that else black gets 3 free moves (g5, f4, g4) for his/her Kingside attack
    58) Ng7 is right, but the point is 1.Ng7! Bxg7 2.h8Q!!! Bxh8 3.Kg6, after which black must watch while white's King wins the bishop and then the f6-pawn.
    60) The solution given is Bg1. I guess it's a good/bad bishop thing
    61) As already seen in the comments, the solution is 1.Bg2! Ke3 2.h4 Kxf4 3.Bf3!! (the point of Bg2). Zugzwang, White wins
    62) I would have liked to see your face after 1.Rg7+! Kh8 2.Bxe5?? Qxc2 3. Rg5+ dxe5… However, 2.Rh7+!! wins: ..Kg8 loses the rook, ..Rxh7 the queen, and ..Kxh7 is illegal
    71) This is INSANELY complicated. The main line is 1.Bd2! Qa4! (1…Bb4? 2.Bxf7+!! Kxf7 3.Qb3+ Kg6 (else the queen is lost) 4.Nh4+ with a quick mate; 1… b4 2. Bxf7+!! Kxf7 3. Qb3+ Kf8 4. Ng5 Nd5 5. exd5 Nb6 (5… Nf6 6. d6 Qd5 7. dxe7+ Rxe7 8. Bxb4) 6. d6 Bxg5 7. Bxg5 Qd5 8. Bxd8 with an extra rook) 2.Bb3! Qxe4 3. Ng5 Qd4 4. Nxf7 Ra8 5. Nxe5+, with an extra pawn and a big attack.
    72) This is a color-swapped Reshevsky v Petrosian (1953 Zurich). Tigran played Re6! sacrificing the exchange, and Re3 is the solution given. It's been debated whether Ra7! (your move) was better. On the one hand, if the rook isn't taken (which is probably better), it looks funny there. On the other hand, who wouldn't take it?!

    Personally I missed a couple of the easier mating attacks, was WAY slower and got a 2200 in the end… Keep up the great work!

  29. "Based on your move choices, our estimate of your Elo rating is 2304, with a 95% confidence interval of [2159…2448]" I"m only 1709 Fide!

  30. I don't trust this, I did this a while ago and got in the mid 2300s while my OTB rating is about 1800. I do believe I am underrated but close to 2400 is obviously absurd

  31. Pretty cool but not sure how much I trust this, I got 1826 but my rating on lichess is 1509…

  32. I got 1891 but I've never played chess seriously, in fact I don't have a fide/other rating and on lichess I average about 1300, so I don't know how this might be accurate

  33. At 15:30, I played Ra1, missing Qh1. 🙂 I finished with 2077 elometer, but mouse-slipped on one problem.

  34. This test is bollocks. I have been playing chess for years and I'm an extremely lowly-rated 1300 player on different sites on the internet. I took this test and it told me my strength is 1901. With 95% confidence I fall between almost 1670 – 1931. Apparently, someone who still hangs queens and major pieces, and enters forced mates in every other game, is near Master level strength according to this "elometer" site.

  35. Anyone getting "Fatal test exception encountered. Test halted." ?

  36. I did the test back in March and got 1877. I don't have elo/national but on lichess my rating hovers around 1300 in 5 minute blitz, though my lichess tactics are around 1855. That's much closer to elometer which has a [1751…2004] 95% confidence interval in any case…

  37. I'm about 2100 elo, and got 2595 on the test, which seems a bit on the generous side. I had fun, though. 🙂

  38. Mine was so off the charts it crashed the site: 'Fatal test exception encountered. Test halted'. I guess it doesn't do sub-zero elo very well!

  39. 34:15 I didn't check with an engine, but I notice you didn't mention h4. I feel like allowing g5 would be far more annoying than giving up the g4 square, and since it's hard for Black to get a piece in there white may be able to play g4 later without sacrificing. Also, the threat of h5, Nh4-g6 is annoying. I don't know if you discounted g5 as tactically flawed, but I can't see it. I don't like c4 as much because d5 can be protected easily; d4 is a worse weakness in the long run since it protects e5.

    46:36 This is a Saemisch variation type position. With Queenside castling I feel white needs to play on the Kingside despite his space disadvantage there. Otherwise there's no point to castling early. White doesn't have an immediate breakthrough on the Queenside. Your suggestion of c5 is dangerous; I doubt Black will take, he'll instead encourage the c-pawn trade because then he gets the nice open c-file to the white king that black's rooks are in better position to take advantage of than whites (how does the KR get to the q-side?) White needs to seize control of the squares on the kingside and then open the g and/or h-files. Qe1 is the right move; it may be thematic for black to trade the dark-squared bishops, but the key factor here is time, and he's taken too much of it to get the bishop to h4; trading this tempo-eater is to white's immediate benefit. As you correctly note blacks q-side development is non-existent and he's therefore overstretched on the k-side. Alternatively if white does want to play on the q-side, then you must save the bishop with Bg1, but I think it's dubious to go charging in to where black's undeveloped pieces already are, opening lines to your own king in the process.

    59:56 Qb2 slightly better. White can kick the black queen out with Ra3 threatening Rc3, and then the Bishop can go to b4.

    1:02:42 This one is tough to evaluate. Even Stockfish thinks a whole ton of moves are plausible. a5 seems to come out of top because it exerts the most pressure on Black's position. Rc1 and surprisingly Kf2 are plausible waiting moves that Ng5 rates inferior to.

    1:19:50 Even tougher than the last. The only real advantages white has are 1) his KR is in play, 2) his K is one step closer to the action than the Black King, and 3) The Black QR is out of position on c5. Any white move must take advantage of these. Rd5 is an interesting try in that after Rxd5 exd5, Black's only good move is Ne7; however, that's not hard to find. I don't see any other truly constructive Rook moves. There also doesn't seem to be a way to make the black QR more uncomfortable than it already is. Therefore Ke3 makes sense, moving it up and defeating Nd4 as a tempo gain. It's still more than likely drawn.

  40. elometer crashed three times on me. I never got further than the first dozen problems.

  41. 55:23 Chess problem 62
    why wouldnt you play Bxe5 first?
    i would say it's better because there you threaten a mate in 2 or his queen gets murdered 😀

  42. I took this a while ago and got 2150 which is where my online slow chess ratings come out at roughly. I had an advantage though because I've gotten to 2200 on chesstempo puzzles and endgames are a speciality because I'm a deep calculator. The one thing though is that the test used to be timed, for non titled players that probably made too big of a difference.

  43. I don't succssed open this website but I stopped the video every time before you answered and I think this positosn were very easy, I answered like you in all the positions ans saw the all varitons except qh1 in 32 and in 35 I found a longer line r:g7 k:g7 qh5 by b:ft if king takes mate in one, b:f6 qg4+ qh8 qe4 and I missed rh8 and in 41 I saw after it kc2 not kc1 but I thinks that this is also winning and but I Bellevue it doesn't meter because if this problems was happend in one of my games I was thinks more and continue with the right moves.
    I have 1900 elo but I believe that I am more than that.
    Edit: when I continued I found that I had a some solutions that not exactly like you but we think about the same things, just decided different so I checked this moves and after all when I compare yours solutions and mine my result is the same like you, maybe i have 30 and this is very surprising because you are like 800 elo above me.

  44. Wow you found in 42 c5 b:c5 bb5 Cy bc4 so aster then me but I was faster in most of the positions.
    But I think that you missed that on ba6 white have b:a6 r:a6 qd3! with a fork on the rook and the knight

  45. In 49 I saw kn:e6 but I also saw first R:e6 f:e6 bg6+ with the same idea, I think that this is better because maybe after b:f2+ qf6+ knf4 the win isnt so clear

Leave a Reply

Your email address will not be published.