I understand that it is a gambler's fallacy . Maybe my ideas aren't right. Maybe the bot is just stupid so are the algorithms. But, I have got to try it so that one day I will realise that it was 100% failure. If by any chance (at least 1% probability) it isn't stupid, then maybe it might be useful. I am trying for that 1% chance of success.

In a normal scenario, when we play it is in no way possible to predict the spots. But, the bot has a special way to learn the spots and predict them in the next sessions. The predictions aren't 100% accurate. But, the bot reduces the stake when it is below a certain percentage of accuracy prediction.

Consider today's test:

https://www.satoshidice.com//user/f05f838b104da3073fc7572f03c3ad8e

The bot lowered the stake to 135 BTC when it had a lower accuracy prediction.

The bot is now in the learning phase and as such is supposed to make mistakes. Only then can it improve. But at the end of today's session, it was again a profit of 130%.

I agree that there aren't any dice site with 0% house edge. Consider the long term period over which the player will definitely hit the loss to be a straight line of 100 as follows:

0         20      40       60       80      100
|_____|_____|_____|_____|_____|

There is a possibilty that somewhere in that line the player will be able to hit successive wins for a particular number of rolls. Maybe from 20 to 40 or maybe 60 to 80. The bot is supposed to analyse which are these slots and learn how to predict such spots. I am in no way making any claims. The bot is still at its learning stage.

Currently the tests are being run to determine the average number of rolls for a session against "Win Chance" greater than or equal to 65%.

We too aren't sure how the system will behave for a large number of sessions in long term.

Albert Einstein once said There is not the slightest indication that nuclear energy will ever be obtainable. It would mean that the atom would have to be shattered at will.”. A year later, Enrico Fermi discovered the way and proved that Einstein was wrong.

My bot isn't about magic. It is created to provide the profits when it is run for a particular set of iterations for a particular "Win Chance". Of course conditional probablity is one part of the system but it is used to make the prediction model better as a supervised learning system. The idea is to remove the "luck" factor for a limited number of rolls (end of a session).


 - Nope. No plans for selling it.
- Except for two algorithms, the rest will be made opensource. Anyone can check for code themselves for a virus.

The bot has 7 different algorithms that are designed to work with different ways. During the beta tests (which is still continuing), the succession of wins were 15 on average with a profit of 12% on average. These wins were with the worst case settings with 100% reinvestment. The idea is not to run the bot for a long term. The basic idea is to run the bot in sessions. Each session will have a limit to the number of rolls.

With the worst case scenario setting (with 100% reinvestment), a session with 10 rolls will lead to 120% profit.

The testing is done now using the 200 satoshi base value offered by satoshidice itself. In the next stage of beta testing, we will use larger investment.

-WTakeDown Team

Hi Guys,

We are the team behind WTakeDown Bot, a Satoshi Dice bot that is created based on our own proprietary algorithm. The bot use artificial intelligence to make the profit percentage to a high one when it is sure that there won't be any loss. We have been testing the bot with 100% reinvestment settings (total amount is made the bet amount every iteration) to know the extend at which the bot can survive a particular scenario.

https://www.satoshidice.com//user/df7d8e704fd69e9e72a09a7227046e7f

Of course, it would be better to set the bet amount to 35% to 40% of the total in each iteration.

The beta testing is going on and we will keep you guys updated. BTCBTC

Once the current tests are over, we will do the ideal settings test which involves 35% reinvestment.

-WTakeDownBot Team