Will investors get Bitcoin Gold later this month?
Personal opinion below. Do not mistake as official statement. No guarantee about accuracy of the following.Ok, so the situation around "Bitcoin Gold" is currently rather blurry. It seems some details aren't fixed yet and might change in the upcoming days before the 25th.
There's a lot of confusion I sense in the community and a lack of necessary information/communication from the team behind the attempted fork.
What I could gather so far is very basic information and it's conflicting in several aspects. The static page with no details besides the date isn't helping to shed any light on the project either.
Going back to the start, Bitcoin Gold was initially announced as BitcoinGPU/BTCGPU (the btcgpu.org domain indicates that). This happened as far back as July 29th[1].
The website back then was a copy of BitcoinABC's website (which was promoting Bitcoin Cash at that time).
Back then, BTCGPU planned to fork immediately after BCC/BCH had forked, something that (as we know by now) did not happen.
Another interesting part of the website back then is the mention of plans about an
ICO. The website stated that "BTCGPU will create 16000 blocks after forked from No 478558th Blcock."[2]
In the time after this, I've lost track of the BTCGPU project (which was since rebranded to Bitcoin Gold, although domains and the git remain named BTCGPU).
This brings us to the current status of the website and the information available.
Their website (
http://btcgpu.org/) is a simple static page with a logo and a date (October 25), as well as some social media links, and a link to github.
Now after having a look around their git:
In September code got merged that would fork BTCGPU nodes (& miners) off the main bitcoin network with block #487427, which they estimated to happen on October 1st, but went down a bit sooner[3].
This code would result in a 16000 block premine at the point of release (due to minimal difficulty and other factors), and appears to currently still be a part of their master branch and available client.
If this code was executed, it also means transactions after block 487427 are void for Bitcoin Gold and website snapshots taken at the point of release are inaccurate compared to actual wallet balances.
I'm not running a BTCGPU node, so I cannot verify this, but it seems although the code was merged, nodes did not execute it and are still on the main (bitcoin) chain as of now.
At least this is the claim of one of their core contributors. Also, he states, "that a premining is being discussed and if there is one, it will be completely transparent to everyone."[3]
How this "completely transparent" premine would look like, how many blocks, under which circumstances, all these things are unknown at this point.
Another thing currently lacking in their code is two way replay protection,
something essential for any fork to avoid replay attacks and problems with transactions down the road (which would include distributions by CG and other sites).
The current status is, replay protection is a WIP right now, and will be done before the launch, which would be postponed if it's not ready in time[4].
Taking all these things together, there's still a lot of if's, when's and maybe's. Some of those might change heavily and rapidly in the next days (at least this is my assessment).
I wouldn't expect an official announcement from CG before all important aspects are fully cleared.
This includes a firm forking date (block or timestamp), a binding statement about any premine, as well as working replay protection.
As long as these things aren't checked out, the decision about if and how CG proceeds regarding BTCGPU might still change.
I'd expect them to handle the fork similar to Bitcoin Cash, and proceed with distributions,
as long as the circumstances allow it and a distribution seems reasonable.
[1]
https://bitcointalksearch.org/topic/btcgpu-change-sha256-to-equihash-gpu-mining-only-protect-decentralize-2046790[2]
https://bitcointalksearch.org/topic/ann-bitcoingpu-btcgpu-forked-spinoff-new-pow-adjusting-size-cap-segwit-2133536[3]
https://github.com/BTCGPU/BTCGPU/pull/2[4]
https://github.com/BTCGPU/BTCGPU/issues/18
Three weeks seems like it would be enough time though. 
