I believe the script is a good starting ground to the bot, but there needs to be a few alterations to make it considered undetectable, and i will list the downsides to each below.
But before i do, we need to come to a realization of a few things, that even though Rich bailed on us. There is still a chance we can make this work, although its very small, it still exits. I have knowledge of CS, but have neither the time to learn all this, or even manage this whole thing. Also, if we do attempt on making the bot actually work, we would need to find someone discreet, knowledgeable, and TRUSTWORTHY; He/She would manage on doing a few alterations to the bot, to actually make a few changes to make the accounts appear more normal and realistic.
This will lead to an increase in monthly cost, and a decrease in potential earnings, but it can mean that the bot can actually survive, and payout for a long period of time in more discrete manner. If this is done right, that means you can work hard to set it up and sit and relax on auto cruise control with actually getting paid for a long time, undetected. The proposed changes are posted near the end of this post.
I think the inherent problem with the bot that all accounts are somewhat too similar, along with the services everyone uses and such. If we can tone things down a bit, then we can make it much much harder to detect that it is actually a bot rather than an account.
The whole idea would be that all of us who bought an account would pool together and create a sort of a voting system. Each script purchases is a equal to one vote. Every decision is put through a vote, for the sake of transparency among us all. A Pooling system with all of us working together means that we all have a chance of making this work, rather than just accepting the fact that we were scammed and move on. Because the script works, and that means that we just have to make sure that we make the accounts look and act more like humans, they cant shut down the accounts if they cant distinctly find out whos a bot and whos not.
We could post up a thread on services on this forum looking for a developer who can help, we screen through candidates, and we vote for the appropriate ones. The key idea here is that all of us search on online sources for Trustworthy and Discreet developer(s) who have knowledge on how to make the appropriate changes. In return, we could offer a chance to run his own copy of the bot. A win-win. To take this further, we can ask the developer to reset our VPS OS, reinstall the bot, so we have a clean run of everything. We could pool a bit of btc as compensation between us (I saw he almost scammed $3100 dollars, an extra few hundred dollars between all of us for everything to work, i'm willing to chip in). So we search for and choose a developer, and then decide on compensation package (we pool money, if any), and then the developer makes the appropriate changes to the script and bot, and we run everything again.
For this to work, someone among us would have to choose to manage the whole plan, and keep everyone updated as a group, and have everything to a vote. Ofcourse, this person would have to be trustworthy, and all money managed (if any) is always kept with a reputable escrow we decide on. This is Key, and make sure that since the money is with an escrow, it is secure from any single individual trying to steal more.
Like i said before, i dont have time to manage this whole thing, and to make it more transparent that this whole idea is written out of good faith, rather than an attempt to further scam, i wont apply to manage the whole thing. But if whoever is interested writes why he/she might be the good leader to run this, please write and tell us why, and we would then take everything to a vote, and choose who runs the new plan to fix the bot.
I will be more than happy to speak to the developer with ideas, and come here occasionally to help out whenever i can, and be a member of the voting system.
These are my ideas, on what we could do to make the bot better and the whole process more undetectable:
1. From each IP service, a maximum of 10 IP's are to be chosen, these are to be dedicated IP's (and preferably virgin). This means that things are a bit more spread out, in terms of randomness.
Downside: the hassle of buying from more providers, and managing everything.
2. For Captchas, A switch from the AI model (thats cheaper but can be wrong sometimes up to 6 times in a row) to the human-based one means that your accounts are much less likely to be flagged for continuously failing captcha's. I dont know if freebitco.in or the doge equivalent actually have any flags for this; but if they at least record how many captcha's you fail, our accounts would be on the high end of the spectrum with a rather bad ratio of number of times captcha solved / number of times captcha failed. That could alone, be a measure of how they could detect accounts.
Downside: You will increase the cost of the AI service, to almost $100 a month for 120 accounts, from around $20. Cost vs Reliability Here.
3. Instead of having a standard sleeping schedule for each account, you could have it set at random for each day. For example, we could have the bot have this sort of dynamic sleeping schedule. For each day, it chooses a number between 13-18, this represents the number of hours the bot will sleep after it has 'awoken'. Example, the Bot 'Woke up' at 02:00, it rolls 15 hours, and thus it will set 02:00 + 15 hours = 17:00 as the sleeping time.
Downside: I think none, other than a slight volatility on the number of rolls each done total each day.
4. Lets take the idea from number 3 further, rather than have the accounts roll each hour for the time they are awake. We choose a 3 hours a day for each bot (or even more to appear more realistic), and in those hours the bot does not roll for any earnings. This is to even further have the accounts appear normal without having too long of consecutive periods rolling.
Downside: You will decrease the number of rolls in the day by almost 1/5th (in example), thus reducing possible earnings.
5. This might be a stretch, but for the sake of ideas, lets put it out there. For each account, choose 3-5 different days in the week (make sure that each account has its own dates, randomly chosen for the month, different from other accounts). This means that the accounts are not always rolling every single day of the year, mimicking human behaviour.
Downside: Again, decreasing the number of total rolls, and reducing possible earnings.
6. Make sure that emails generated are from all different providers (not to mean that they wont be overlaps, but that it would be odd if 120 accounts created all have yahoo emails addresses. A 1/5th split (or even smaller) between outlook, gmail, yahoo, aol, etc. makes everything appear more realistic.
Downside: None
7. The overall hierarchy of account referrals should be somewhat random. Dont make all accounts follow one person, keeping changing up which accounts refer which sub-accounts. Make sure that one account only refers a maximum of around 9 different accounts, and any of the 9 account do not refer any more people. That way, you always have batches of 10 accounts linked to each other and not more. each batch does not link to the other batch of accounts in terms of referral hierarchy. Also, maybe spread out creation of the accounts over a a week or so.
This is what i can come up with so far, and more ideas could be suggested. Let me know what you guys think, and feel free to bounce any ideas around.
Edit: a few ideas added