Since 1st January this year there have been over 1000 failed attempts to access pages on my trains website. This site is regularly checked using a spider application and, apart from five outstanding minor internal targets, is error free.
The main reasons why users receive 404 error reports come down to:

• Incorrectly typed URLs. Remember that web addresses are case sensitive. "br103.htm" will not open "BR103.htm".
• Renamed pages. E.G. BR103.htm has changed to BR103.1.htm as BR 103 has several sub-classes. I still have people trying to access files that have not existed for a couple of years.
• Incorrect deep links from other websites. These are often linked to the above reasons.
• Repeat visitors often go to the wrong address after a page address has changed as the old address is in their browser's cache. Not much I can do about that except to suggest reducing the size of the cache. (Mine is 0 – no cache. It solves a lot of problems during editing.)

There are other reasons (e.g. web searches often throw up some weird addresses, some that only existed for a few hours.
So, if you hit a 404, try going back to the home page, or one of the index pages for major zones (N, Z, Layouts, layout, etc.).