For the statistics, I added a virtual bank of 10,000💰 to better track its verdicts and understand where it makes mistakes. The result: if it had traded with its virtual hands, it would have made +700💰 yesterday, but don’t think that’s a lot, it's just 7% of the bank. Today at lunch, its statistics were -340💰🤣🤣🤣
He probably turned on the gambler mode🤣🤣🤣 At the moment, the loss is already less than -130💰 for the day. He has recovered more than half🤣🤣🤣
But I'm curious to see what he thinks, look at $SIREN , yesterday he managed to build a trend line as we observed a slight decline, but look at his verdict🤔
At 4 hours, his long forecast was 75%😳
At 1 hour, his forecast was for a Long 55%😳
At 15 minutes, the probability of a short was 55%
Conclusion: in the long term, he saw a Long, but there was no specificity on the 1 hour and 15 minutes. Of course, if we had entered from the current position, a stop-loss would have triggered, and we would have been slightly in the red, but his verdict is vague, he didn't see a Long in the near future.
And today I was curious if the price would reach his take after such a powerful decline❓ And what do you think❓ The price reached, even slightly surpassed 2.875, although he predicted 2.80.
This is interesting and it really helps, but it's not a magic button, if we had waited an hour (and that's the time when he couldn't decide Long or short) we would have seen the collapse, but exactly after 4 hours, we could have looked for a point to enter Long, but that would have been scary, then we could have set a notification above the flat and jumped on this train. Although, maybe it's just a coincidence❓
I was also attracted by $M
The photo on the left arrived about 6 hours ago, it's kind of boring, he was only 60% sure about the long, but look what happened after 2 hours😳
On the daily, Long 💯 %
On the four-hour Long 💯 %
At the hourly Long 62%
At 15 minutes Long 75% 😳
I certainly didn't understand, did he believe in himself or what🤣 But I see such a combination for the first time, and in the second photo I told the AI not to be lazy, as he was missing indicators, and to show me how many indicators were for Long and how many for short. That's why you see a different format of messages.
And just so you don't think everything is so rosy, I'll show you the biggest and dumbest failure, the AI said to take Long and since then the coin $ONT has fallen by 17%🤣
Thinking with my head, it's 80% short here, even if you look at the overbought RSI on all timeframes. But as I said, I changed the request structure to the AI today so that it looks at each indicator and makes a verdict based on the pros and cons. Although as you can see he suggested a safer entry at the level of 0.06, but if we had entered at the current price, we would have lost -17% by this moment. I also took this into account, and now he should choose a justified stop-loss no further than 5% from the entry. Let's see.
Still testing models, their accuracy also depends on the AI's decision.
I read an article that AI was given 10,000💰, GPT and Gemini are in the losers, while DeepSpeek nearly doubled its capital in a week. Something on coinglass isn't loading AI statistics today, I can't verify the accuracy, more precisely, I know that GPT has been in the losers for a long time, but I'm curious about DeepSpeek.
Today I'm testing a new model from GPT, it's cheap and top 5 in finance.
I want to find something that is both inexpensive and smart🤣 Those who are connected with AI will understand me, I have 700k tokens per day for entry and 1.3 million for exit. If I take the smartest models, the daily cost will be 20+💰 But DeepSpeek will be relatively affordable🤣
For now, I'm observing, until today there was a free model from stepfun, by the way, it's good, but they say free models are lazy🤣
@Binance Square Official Can we somehow add the ability to send photos to your new skill for posting, and not just text? That would be great.
#AIBinance #BinanceSquare #siren #clowdbot #OPENCLAW
