Performance Evaluation.

Performance Evaluation.
Lichess.org offers basic stats, which is really not much; but in fact it does a LOT more than ANY other chess web site would do without monthly fee. The question is how to take the maximum out of it.

It is commonly recommended that your opponents are slightly stronger than you. If you choose the opponents who are a little stronger than you then you can achieve the best progress and learn the most from your games. Under this premise you can analyse your basic stats judging your level depending on how many games you win or lose.

So this is what you can do: find the numbers of games played on your profile. Use the number of games won and divide it by the number of all games played. In my case it would be (631/1159)x100=54% ! Which means that I pick opponents slightly weaker than I ;) Truly I am not that kind of person and I can explain why this happens: it is relatively difficult for me to find an opponent rated at 2200, because sometimes I’d need to wait 5 to 10 minutes before the game begins. That’s why I set formula range at up 2000 elo. Then someone comes immediately. And the folks around 2100 are also quite challenging for me. I am satisfied playing that level.

Anyway, watch your numbers. If your win ratio falls at 45 – 50 % - it’s excellent! That means you are doing your best. 50 – 55 % is also okay. That means your rating is well calibrated and you play at your level. If your win ratio is about 60 % - this could mean that you pick those who are weaker than you, that your elo is boosted (inflated) and you don’t learn from your games as much as you could. Similar - if your ratio is like 40% - not good. We are all humans, we can’t become GMs by simply playing with GMs.

Another useful piece of information would be the opponents average ELO, and they had this on lichess in the past, but the developers removed it for some reason, I have no idea why. But you can find your “favourite opponents’” ratings on your profile. And that can remind you of who you play.

I systemised and added new possible factors of abnormally poor performance (bullet) following this discussion:
lichess.org/forum/team-second-life-chess-community/bad-conduct?page=3#22
Only those effects are revealed which have nothing to do with physical condition or disadvantage of any sort (lack of sleep...) I hope this can help to prevent much of improper suspect.

1. She is not experienced at bullet and possibly encountered a “bullet nerd”. “Bullet nerd” type: very experienced, plays only bullet every day around the clock since the age of 8 years old. Characterised by the extreme mouse dexterity (very critical for 1 min bullet!) Knows lines. Often plays the same opening lines with both colours.

2. She lost because she played not her time control she specialized at: 1 min game will be won by 1 min player vs 2 min player. 2 min game will be won by 2 min player vs 1 min player.

3. She lost many points because of the “lotto effect”. If she normally wins 8 games out of 10 vs a 1600 rated player at blitz, she will win maybe 6 out of 10 at bullet vs the same player. Bullet randomises the outcome and increases probability to lose vs. a weaker player and lose more rating.

4. Higher volatility at bullet means lower deeps and higher ups of your score unless you stop on the way up at some satisfying comfortable score. Normally strings of good luck should offset strings of bad luck and catapult your rating much higher. But the feeling of uncertainty motivates players to stop when a certain satisfying level is achieved. That’s why they often may be having lower bullet rating than blitz rating!

5. She lost because her opponent was underrated:

5.1. her opp. was a very strong player who recently joined the site (or just started playing this variant) and was about to gain higher rating;

5.2. her opp. was on the bottom of her volatility range for any of the mentioned reasons;

5.3. she lost because she played long bullet time control (ex. 1+3, which resembles blitz) vs a player who is very strong and highly rated at blitz;

5.4. her opponent cheated;

5.5. the opp. was “bulked”. This is an interesting phenomenon pertaining to online blitz a. bullet chess. When 2 equally strong IMs play online only with each other starting at 1500 elo then they will be fluctuating around 1500. In that sense they will be sort of bulked. This theorem can be spread on more players involved. It can be 100 players or more, whose “real” rating could be like 2200, but they would be chronically rated at 1900-2000. Once you get in such pool of players it’s like a morass that keeps you there, and it’s very hard to escape.

6. She was underrated because she was bulked herself.