
A Preliminary Analysis and Brainstorming Blog of Potential Reasons for Chess.com Crashes
Background
I think everyone at chess.com has felt frustration with the current problems, overloaded databases, endless repetition of the same information, and other problems. We know that the staff at chess.com are putting all their resources into solving this problem, but we, as a community, can also help.
Problem Definition and Conceptual Framework
"A problem well-defined is a problem half solved."[1]
The general problem is the increased crashes because of system overload. It is a complex problem or a "Mess" as defined in creative problem solving "CPS." In CPS, a mess is a technical term that describes a cluster of interrelated issues, each of which must be considered by itself to address its larger problem.
CPS is a well-defined process comprised of divergent and convergent thinking. The divergent phase of CPS consists of identifying potential solution paths without judgment. It is essentially a brainstorming session.
Once paths have been identified, the courses are evaluated in the divergent phase, reducing the number to two to three options. Research indicates that exploring all options before choosing a possible solution path Is advantageous in solving novel problems.
Applying CPS to the chess.com problem
Let's start by looking at the observational data.
DATA
- A slowing down of processing speed.
- Frequent notices that the databases are overloaded.
- Increased spam-like glitches, repeating the same message over and over again without an option to escape or delete
- Glitches spreading to related areas, such as a glitch in the tournament announcements, repeating the same announcement repeatedly, and blocking access to playing individual games.
- The Inability to use the message system during certain times
- Spam-like repetitions of the same announcement on the home page cannot be deleted manually.
Other observations: I noticed that doing 8:00 a.m. and mid-afternoon Pacific Time, the problems appear to be at their worst. Towards the evening, the problems seem less severe, except for the spam-like loop of announcing arena tournaments.
Key Findings
So, let's summarize the key observations and look for potential reasons for these occurrences.
- In the initial phases of this problem, most of the issues occurred between morning and mid to late afternoon Pacific Time.
- Over time, the problems increased in severity and spilled over to other time slots. For example, the repeated tournament announcement problem does not seem to be who sticks restricted to a specific time.
Interpreting the Findings
Let's first look at problem #1. What would increase the problems during a given time range? Two potential "candidate ideas" are:
- A time-based increase of activities that are resource intensive.
- A time-based sudden increase in participation
Nakamura's Analysis
In his YouTube, Nakamura discusses the crashes on chess.com.[3] He points out the influx of newcomers to chess.com after Twitch's partnership and PogChamps. PogChamps is a streaming program in which 10 of the most popular Twitch streamers started streaming on chess.com, with chess personalities such as Nakamura providing commentary. This increased the interest in chess and chess Utubes and added many players to chess.com. According to Nakamura, this trend continues and strains the servers.
In summary, the increased focus on streaming, TV shows, and other bandwidth-intensive activities, in addition to the influx of newcomers, may have strained service and nearly pushed them to their limits.
This happened in 2020, and PogChamps has continued to be popular and brings increasing numbers of participants to chess.com. However, it seems insufficient to explain the sudden extensive problems chess.com has been experiencing since the beginning of this year.
Expanding Nakamura's Analysis to the Current Situation
On December 16th, 2022, chess.com acquired NorwayChess24, which most likely resulted in a sudden increase of the online population that pushed the already strained servers to their limits when the waking time zones in different countries overlap. If this is true, there might be ways we as a chess community can work together to reduce the strain on the server until chess.com can fix it.
How the chess Community Might Help
Plan high-bandwidth activities in low-traffic times (this will vary by location).
When scheduling tournaments, consider the target population and see if it is possible to organize these events in a shared, non-prime time zone.
Use a knock-out format for larger group events where the preliminary rounds are organized by country/time zone.
Conclusion
This blog is intended to be a conversation starter of collaboration while chess.com works on fixing the problem. We are one community, and we all have a vested interest in working together. Furthermore, we are chess players and problem solvers. Let's put our community's brain power together. No telling what we can do.
[1] Quote attributed to John Dewey
[3] I am unsure if it was a YouTube or a stream, and I cannot find the reference. I would appreciate it if anyone familiar with this particular YouTube would share the link. It starts with a list of possible reasons and focuses on the abovementioned issues.