Topic: AI prediction/forecasting for Bitcoin and crypto trends - page 2. (Read 843 times)

There are so many AI prediction for Bitcoin prices. As a matter of fact, I came across one project like this a couple of days ago. I have played around with models like this but end up getting confused by their predictions. I think more details needs to the added to the model to make it more robust. You have done a got thing no doubt but I have found that ai models can't make long-term predictions. How are you able to manage the humongous processing power needed to run it?

It is indeed as you say looking at what is presented there in the chats using the key nations but, you keeping an update on this using future predictions would be a better clarifier as, forum users would be on the prowl to see how it performs. If it could give the same results as in OP.

Should the later be true, I wonder what would become of trading as a whole. Especially when it goes down  to the fact that, many would want to depend on this for analysis and market prediction. Maybe we would have an era change where users no longer want to use indicators no more.

This is incredible. Price was ~1% up 30 minutes before closing, and then it dropped to -0.28% to close the day, making prediction correct. There must be some hidden order behind.

Prediction for 30.07.2022 is downwards.

Current standings since start of the predictions:

Prediction for 29.07.2022 is downwards.

Current standing at 9/9.

Daily candle for 26.07.2022 closed at -0.22% at price: 21259.78, as predicted downwards. Price rallied on 27th 8%, and there was no prediction, as I wasn't able to run it that day.

Still standing at perfect accuracy, but expecting it to fail soon.

I think your omission of the 10-15% upwards movement, across the crypto spectrum, between 12am (EST) 7/26/22/ to 12am 7/27/22 would be considered a strike, a costly one, but I’m still rootin for you.

Prediction for 28.07.2022 is upwards.

We will have to wait for upwards movement for now. FED is meeting today and tomorrow, probably going to increase interest rate, and markets might go down heavily. Good thing AI agree on all of it.
Current standings: 6/6.

Prediction for 26.07.2022 is downwards.

Keep'em comin nites, my retirement is on the line. Soon as U say upwards , I'm placin some baller orders, and sayin bye to day tradin, when the $$$ rain down on me:-)

This one will close downwards as predicted, as BTC dropped more than -4.00%.

Wondering if this guy above is shocked, since I must stole something from people that control market.

Next forecast will be tonight when candle closes.

Prediction for 25.07.2022 is downwards.

Another correct prediction, currently standing at 5/5 accuracy predicting direction of daily closing candles.

I will not be able to run the system for tomorrow, so the next forecast will be for Monday, 25th.

Prediction for 23.07.2022 is downwards.

It is machine learning system. Machine learning is a subset of AI. I'm not using any technical indicators or macro-economic indicators in the system.
It's using logic and algorithms from speech recognition engineering. Price is processed using various algorithms and then machine learns dependencies, patterns, basically learning how to predict it.

I will continue to post daily predictions. Let's see the result for e.g. 30 days.

Current standings:


Still can't append photo for some reason.

What is the basics of this? Is this just using multiple indicators and price action? Or is it factoring in all the macro economic indicators as well? I suggest you give some thesis behind this if you want someone to believe this idea. To me so far this looks imaginary let's wait for few days to see if the market still moves as per your line. But will love to know about this in detail.

Let me break it down for you, as I'm not yet sure if your brain is not capable of understanding simple things or you are just here to cure your frustration from not having any knowledge about the topic.

When you train and test AI, you split the data into the training and test phase in the dataset. Test phase is where you test system performance. Then you can use that to predict the future.

I posted forecast on July 15th:



and then 3 days after I noticed I didnt made forecast but these predictions are the last 4 days in the TEST PHASE from dataset.



What you see here is prediction for the test phase, not the future dates you are so stubbornly trying to point as failed prediction. And I OPENLY said it was my mistake and it is not a FUTURE forecast.

And how many times I said in this thread I'm not trying to forecast exact prices, but the trend and direction, and somehow you failed to even read that.

After that I said also multiple times that I will forecast one day in advance until I fix multidays forecast, exactly what I am doing last 4 days. Because you can't read the thread, here's quotes of them for you.


Now take that, and compare it where market closed.

Once again, if you are going to notoriusly lie and take all the things out of context, please restrain yourself from commenting.
You have made cool joke, expressed your frustration and it is time to do that on the next thread, as it seems that's what are you doing in your life.

If you are somehow able to write intelligent comments like other guys here, I would be very happy to answer you.

Yes, I do!


Again how do you call guys that claim they have a tested system with 90% accuracy that flops from the first try?


Can't you read properly?
I already told you my problem is not with your failed prediction as I already knew those numbers will not be achievable ever, but with the fact that you quoted a correlation with the real numbers on the market which were definitely fake!
You can go back to your post where you claim that

and tell me where you got the "real prices" of 19970.56 and 19970.56 since it never went below 20k since mid-14th.

Can you read properly?

This is what I wrote:

It was prediction in the test, not the future dates you said. I stated that in few posts. Why are u taking it out of context?



I meant on these daily predictions I post everyday.

If you are going to take everything out of context and just lie, I would like you to leave the thread.

Yeah just how some authors make millions selling lottery earning schemes or how some Kiyosaki made millions selling books giving financial advice when in reality he himself went bankrupt twice. Do you know what people call those?


First, don't talk bad about Mr. Goxx, that hamster had a proven record, documented, and he has made real money with real transactions.
You said you were being able to predict 5 days in advance, your first prediction was not only off but you came also with fake numbers for the assuming real price..


This is the data for the days of 16-19


What you claim to have happened never happened.

So since neither the opening, closing lowest, nor highest was anywhere near what you predicted.
You predicted the first day wrong, you went for lower the next one and it got up, and you missed the growth between day 2 and day 4 by a full 100%, going for 1000$ when it went up 2600$. So, it's in the range of hamster accuracy.

Forecasting is a way to take gambling away from the markets. Yes, everyone can make predictions, but everyone will lose money and leave the market if they dont have tools or knowledge to forecast the market with decent accuracy.

Well, analysts are flawed beforehand, they have emotions, greed, they need to pay rent, eat, sleep, etc. That's why I built smart machine to do that. Of course market can change, hence my software is adaptive.


I'm not sure if you are familiar, but if you sell something you get money, and if you sell it a lot of times, you get rich. Crazy, huh?

Maybe it is dreamable for you to trade profitably and make a good predictions, for me it's reality. I predicted last 5 days closing candles in advance, what have you done in that time, except making 'trying to be funny' hamster trader posts.

Edit: well, look at that, today is also going to close as predicted, must be supersymmetry.