Why not fix existing bugs before adding new features?
Hopefully this is a better answer to your question.
The product is still in development, so major new features will continue to be added. Code this complex will never be completely bug free before your start on the next major code change, so you can't wait to start until all bugs are fixed, because there will always be one more bug. At some point you have to say, "OK, it's playable enough, time to add the next big part." While adding the next new complex chunk of features, the code will basically be mostly broken and unplayable until they are complete. Adding these new features also involves changing the code in a way that may make fixing some existing bugs pointless, because they're in a section of code that is going to be thrown out or massively rewritten.
So, you are left with two choices when you start working on new major features/changes: splitting your codebase or periodic bugfixes.
Splitting your codebase means that you are now basically debugging two different programs at once, the upcoming release with the new major code changes and the previous release. A fix for one may or may not work in the other. This means more work, potential problems with mixing up codebases, and a longer release cycle, due to the more complicated process, duplicated and/or wasted effort, and switching back and forth between two different tasks (writing a new feature and bugfixing). The
only up side is that bug fixes come continuously.
On the other hand, periodic bugfixes between releases means that you have one codebase, effort isn't wasted on fixing bugs that were going away anyways, and a faster release cycle. It's a much simpler system. Yes, this means that you get a "new features -> new bugs fixed -> old bugs fixed -> wait during new feature development -> repeat" cycle, which means that the bugs do not get fixed for a while during that third step, but it also means that the product as a whole gets done much more quickly.
Considering that this is a one-man programming operation, the latter is a far more simple and sensible system. While the wait between bugfix cycles can seem a bit long to the testers/players, in the end it means that the game will get completed sooner.
I hope that helps answer your question.