My Journey to 2700+ Live Blitz Rating, a Discussion on Engine-Users, and a Plea to the Community and Administrators of Chess.com
(This was an article I wrote in October 2011 on chess.com)
Live Blitz Chess Stats: Eilyisum
Today’s Rank: #1 of 751, 499
At least at the time of this writing, I am the highest rated live chess blitz player on chess.com with a rating of 2727. Reaching the #1 was an incredibly arduous journey. I dedicated countless hours over a period of more than two months. The whole process not only was a great experience for me, but also allowed me to really get an understanding of the top players on chess.com. I wish to share my many experiences, thoughts, and concerns from this adventure here in this article.
Going from 2200 to 2700+
First of all, if you have access, please follow this link to see my stats page:
Looking at the graph is particularly helpful in terms of understanding my rating progression, but is not absolutely necessary.
I first joined chess.com in April 2011. I quickly established my live rating at the respectable 2200 level. However, I simply could not get higher from there. The reason? Simply put, over half of the players on live chess rated above 2200 cheat. They use computer assistance in their games. These silicon machines easily counter my aggressive and tactical style, especially in blitz. I continued playing bullet chess on this site, where the frequency of engine abuse is much lower, but I was still very discouraged. It ultimately encouraged me to stop playing on this site.
Fast-forward a few months to September 2011. I began offering coaching services on this site and decided to give live blitz chess another shot. Much to my dismay, an abundance of computers still existed past the 2200 rating level. However, this time around I discovered an interesting concept. In some games against engine-users, I would be completely lost (say for example down a queen). Most strong players would play the rest of their moves very quickly holding such an advantage. However, these engine-users would still play slowly, and in fact run out of time in some cases! It then suddenly hit me. I could beat these engine-users on time! I changed my style against engine-users to “survival mode”. Although I still lost many games to engine-users, I also began winning much more on time. My rating started to climb and it soon reached the high 2200s- very low 2300s.
I then had the fortune to play against a human player with an extremely high blitz rating of close to 2600. Although a very strong player in his own right, he achieved this rating by consistently beating highly rated engine-users with carefully formulated anti-engine strategies. He unfortunately could not apply these same strategies against me. In our blitz series, I lost a few but ended up winning the vast majority of our games. This catapulted my rating to 2400 plus for the first time and gave me a huge confidence boost. I made it my goal to achieve 2700+ and become the #1 rated blitz player on the site.
I spent the next two weeks stuck at 2400. I was managing to beat engine-users at essentially a 3:1 ratio, but this is simply not good enough to increase your rating once you hit 2400. To say the least, it was extremely frustrating. However, my computer knowledge was continuously evolving and I was incrementally improving my anti-engine strategies. In my spare time and in between games I would use my own engine to see where I could have improved. I also played practice games against various engines (Rybka, Houdini, Shredder) to learn more about their playing style. In time and after more than a hundred hours of work, I nearly perfected my “hippopotamus” anti-engine opening system. I continued to play my “normal” openings against human players. If you go through my games history and see me play a “normal” opening against an opponent, it means that I at least believed he was human.
My next big rating jump came at the expense of another human player. This player had a 2400+ plus rating when we started our series. Luckily for me, he did not know how to handle my openings very well. I also believed he was just having an off day. In any case, I decisively won most of our games and my rating jumped to over 2500. I rode this huge wave of momentum and followed this up by quickly beating several 2400+ rated engines. Before I knew it, my rating stood at 2600.
From this point on, it was essential that I carefully selected which opponents I played. Losing even one game was a huge setback. For example, a single loss to a 2200 or 2300 rated player would result in a -14 or -12 point loss. Meanwhile, each win against these players gained me only one or two points. Ironically, the most logical opponents for me to target were engines. I can maintain 7:1 or better win ratio against computers with my anti-engine strategies. However, even though I consider myself a good blitz player, it is impossible for me to have similar repeated successes against strong human players.
Even with being very selective about whom I played, increasing my rating from 2600 to 2727 was one of the most grueling and challenging experiences of my life. I could not afford to lose focus. I remember the day I first broke the 2700 barrier; I was ecstatic and let my guard down just for a second. I consequently ended up losing two games. Just like that, in the span of ten minutes, I was back down to 2670. The previous three days of work instantly erased. Luckily, I was eventually able to regain my composure. I steadily climbed to the peak of 2727 over the next few days. I remember the last rated game I played when I was at 2726. I was down a queen and a few moves away from being mated, but my opponent then ran out of time. It was a fitting end to a nightmarish, exciting, nerve-racking, and stressful journey.
Sharing my Engine Detection Knowledge
I am enormously proud of reaching the rating of 2727 and being the #1 rated blitz player on the site. However, I have no delusions of grandeur resulting from my accomplishment. I realize it is only an inflated Internet rating that is arguably undeserved. If I were to only play human blitz players from this point on, it would undoubtedly drop to the 2400 range or lower. I may be #1 rated on the site (note #1 rated does not even mean necessarily best), but there are easily hundreds if not thousands of better blitz players than me in the world. This is why I am not so much proud of the high rating but as to what it represents. It represents countless hours of effort, perseverance, and dedication. I set myself a seemingly impossible goal and achieved it. The most valuable thing I gained from this whole experience is my vastly increased knowledge about engine-users and I wish to share this with the whole chess.com community. Much of the following is already relatively common knowledge, but to those of you who don’t already know I hope you will find it eye-opening.
First of all, what exactly do I mean when I use the terms “engine-user” or someone who uses “computer assistance”? There are two main types of “engine-users”.
1) The first type of “engine-user” is an account that managed to write an algorithm that directly allows the engine to play the moves on chess.com. I am not sure how this is done. In all of my life, I have never managed to beat an engine of this type in blitz. Fortunately, these types of “engine-users” are extremely rare and chess.com is rather apt at detecting and banning them.
2) The second type of “engine-user” is far more common. This account has a commercially available engine running in the background (for example Houdini). This player inputs every move played on chess.com into the engine, sees what move the engine suggests, then makes the recommended move manually on the chess.com board.
How can you tell if you are playing an “engine-user”. Both of these types of “engine-users”, especially the second type, have a few easily identifiable characteristics. I will list a few of them here, but not all of them because I am aware of the fact that this may have some negative consequence of potentially helping the cheaters. However, I believe that the overall increased community awareness outweighs this.
Common Engine-User Characteristics
1) Never makes any obvious tactical errors
2) Monotonous Move Tempo (i.e. 3-4 seconds for EVERY move, despite how complicated the position is, or even for “obvious” moves)
3) Large rating disparity between Blitz/Standard rating and Bullet rating.
4) Ridiculously good win/loss records (i.e. more than a 5:1 win ratio).
With this information in mind, what should you do if you suspect if someone is using engine-assistance? REPORT THEM directly to the staff. Send in tickets. It does no good at all just accusing them to their face or writing bad notes on their profile. You can help the community so much by following these simple steps.
I started out just wanting to become the highest rated blitz player on the site. On my route to accomplishing this goal, I discovered a much deeper and meaningful issue facing chess.com. The amount of engine use at the high level is disturbingly rampant at the high levels in live chess. Some may say this is an issue that only impacts a small minority (.1% of the best players). I completely disagree. Many live chess players enjoy watching the top rated players play, but half the time they are just watching “engine-user” vs. “engine-user”. Engines don’t play like humans, and you don’t learn from watching engines play. Even more importantly, I believe that the high density of “engine-users” greatly discourages the best chess players from playing here. More GMs and IMs would join this website and play on live chess if the population of cheaters was significantly reduced. As a community, don’t you want to have the opportunity to watch these great human players play live?
I implore the community here and the administrative staff to take a more active stance on banning these engine users. As part of the community, please be even more active in reporting suspicious players and further encourage the staff to also take a more active roll. I ask the staff and management here to strongly consider hiring/appointing experienced engine-detecting human moderators to specifically overlook live chess. These moderators should have the power to swiftlyban cheaters if they so see fit. I understand chess.com already has an automated-statistically-based cheat detection system that works quite well for turn-based chess. However, implementing experienced and knowledgeable human moderators who perhaps also have access to this automated detection system would bring this level of quality and cleanness to the live chess section. Cleaning up live chess has many advantages. A cleaner environment would encourage many more titled players to play live chess here. This would be great for the site and even more so for the community. There is something about watching great players play live that simply can’t be captured in the turn-based games.