I can explain this... if you compare the prices shown at bitcoinwisdom and the prices in the software then you will see a difference, this is the cause of the difference in the readings. The software reads the prices on its own moment. As you know the price is never the same, this changes every nanosecond. So here is the base of the difference in output.

I do not understand the focus on re-creating those charts. Personally i do not want this because i do not want to trade at the same time as all the other bots that use those prices. And besides this... what is this software really about? Yes, getting some good trades from it so why not focus on a good strategy?

I know what comes next... who do you find a good strategy? I was just working on a response explaining this better.


If you read this kind of values then it means the response is to fast. You need to slow it down to get some better readings, there are several ways to make this happen.

1. You can increase the update timer of the indicator. (its 2min by default for MACD, so you can have a try to see how 4min preforms)
2. You can slow down the moving averages, for example increase the short moving average by +2. (by default its 12, try 14 for example)
3. You can slow down the signal line of the MACD, for example increase it by +2. (by default its 12, try 14 for example)

Trying these things and watching the effect (and learning) from it is very useful. There is even a way to do this more easy. I had a little hard time to describe it, so i will explain this with a image.



As you see i have setup a trade-bot with a lot of the same indicators to it. But all those indicators have a slight difference. When i now benchmark/backtest this trade-bot it will not do good in trading of course. This is true overkill. But it gives me (and you) the output of all the indicator data and i am now able to really see the effect of the changes. The only thing i need to do is compare the response charts at the "Edit indicator" -page and i am able to make conclusions. (aka improve my strategy)



I really recommend to do these kind of test setups because it just make things more clear.

Of course the general recommendation remains standing: to avoid losses use the "FeeOvercome" insurance.

BTW: The backtest is designed to be pessimistic. If you did a 1 tot 1 compare to a bot that has run then the bot that has run will have made more profitable trades. This is because market orders are used and because real sell prices are used instead of static ones. We will shave this down a bit in the next update. But this concept itself remains standing, the end-benefit of having an pessimistic backtest/benchmark is to large. (to explain this easy: the exacter the backtest would be, the more can go wrong)