There are two many updates ... does that really make sense?

[Frank Quisinsky]

Here my diary message 060:

d060.
July 09th, 2026 (18:45):
At first, you might be happy, but if you think about it logically ...

Once again, the question of purpose arises. Starting around July 22th, 2026, 10 of
the “Super Generals” will compete one after another. Both release versions and
development versions are available. For some of these engines, there are even
be both release and development versions available. One day after a release follow
a dev version?? Often to fast releases, after a release version follow the bugfix
versions.

***
It is likely that all 10 engines combined will have improved by no more than 60-70
Elo. That's my tip!!
***

And that is exactly the point!?

Not to mention that for quite some time now, there hasn’t been a single person in
the world who can win a game against an engine rated higher than 3000 Elo. We’re
talking about engines (these Generals) rated around 3500 Elo or higher.

Why are these top engines updated so often? I see absolutely no point in it, which
once again raises the question of their purpose. Personally, I find every update to
a 3000 Elo engine much more interesting than an update to one of the many "Generals".

Once the programmers have reached around 3450 Elo, there should actually be far
fewer updates. Ultimately, Stockfish will probably remain our number one. After all,
the many “Generals” are, miraculously, very close behind it.

This looks very strange in every engine ranking and is also one reason why many
computer chess fans seem to be slowly losing interest.

Without downplaying the work being done on the TOP engines, if there were just one
update per year, the community would look forward to it more, all these updates are
definitely not a good thing. Thats my personal opinion about it.

[Frank Quisinsky]

I don't see many really good reviews here anymore. Instead, they're often just “praise from best friends.” I think it’s really a shame to see how the community has changed in recent years. About 15 years ago, hundreds of people were organizing tournaments and put the results in chess fora like this. Today, only a few remain who still do that. Mostly people who fall asleep and wake up again with their engines. Right now, I’m the best example of that (not for ever, currently). But it’s time to ask what the point of it all is.

At first:
Are all these updates really necessary?
If yes ... give me a tip to understand it better!

As a developer, don't you agree that you should give the few remaining testers a little time to work through a release?

Don't you think that fans of your engine would be much happier if there were fewer updates? Maybe then they'd test more thoroughly, rather than just skimming the surface as they tend to do now with so many updates?

Thinking on it before you go into an unnecessary argument.

[Modern Times]

I disagree with all of this. Your assumption is that testers are the main or the only audience for engine developers. They aren't. Engine developers want people to actually use their engines. Of course it is a competitive landscape and developers like to see where they stand in the pecking order, but most of them know that anyway with their own testing.

Testers should not stifle innovation by dictating how often developers should release. It is entirely up to the developers, and in fact I'd encourage them to release as often as the like if they feel excited about their changes and want to release a new version.

Testers do not have to test every new version. They can just skip a release now and then. Simple.

[Frank Quisinsky]

It’s not about telling people what to do. Most of the time, people who are deeply immersed in a topic don’t even look left or right anymore. I’ve often felt that way myself, back when I spent a lot of time on Arena or Winboard or doing other things. I had those blinders on. In life, it’s always important to question things.

We shouldn’t always head straight for confrontation; instead, we should take the time to think things through calmly.

For example, if there’s still a battle for the top spots at 3500 Elo, one that’s decided by barely measurable margins of just a few Elo points, then that’s not a valid argument. Generating random ranking positions very quickly before the next update is released ...

Rather, the point is how we can get as many people as before to take an interest in this unprecedented diversity. Perhaps by highlighting new and interesting perspectives.

Unfortunately, I don’t see a single valid counterargument to the situation described in your reply. Please don’t be mad at me!

I’ve written this before, and I’d like to do so again:
Think about it first before getting into an unnecessary argument.

[Frank Quisinsky]

Ah I forgort,

what you wrote to the testers are mostly right.

But that is also not a point:
To skip a version or five versions if a tester like to do it is not the topic.

Personally, I can't bring myself to accept it when I've put work into something, am proud of it, and then see that it's already outdated by the next morning.

If I am thinking on the past:
Months of dicussions about a Fritz version in fora.
And christmas time ... a new Fritz will be out.

Strengths and weaknesses fit together like pieces of a puzzle, because many opinions are always better than individual impressions.

Expectations are very high these days. When an engine like Fritz 21 hits the market, one that has significantly improved its playing style and is quite interesting, the average user wonders …

“Where does Fritz 21 rank?”
The answer: “25th place… The answer: That’s terrible.”
We are speaking from humans!

Yet Fritz 21 is only 60 Elo points behind Stockfish.
If this result had been achieved before the pandemic, the community would have cheered.

Rankings are like neural networks.
They are both a curse and a blessing. The curse, however, leads to a decline in interest in our wonderful hobby.

Because our own arguments are becoming increasingly questionable.