ANALYSIS
Chat GPT-4 and Bard competing for the most plausible price prediction for Bitcoin and $WBT altcoin
Paul Osadchuk · Follow
Published in ·
--
AI has been slinging the raft of big tech world for quite a long time. Setting his foot into technology dimension and our routines, it is seemingly a powerhouse of any knowledge a person is curious enough to obtain, as well as a key to facilitating the basic intellectual tasks.
Artificial Intelligence has tangibly impacted all walks of life and work. Of all the mainstream and niche spheres, it became the subject of scrutiny for crypto investors as well. The technology sought its utility namely in crypto trading bots — AI-based tools, served for expediting the investment strategy by analysing profitable decisions.
Still, trading bots could not be called fully artificial, as they are limited to a certain type of knowledge with human interference. The universal AI models — quite another matter, by contrast. The ones are not confined to a certain topic, or selection of resources, but instead designed for being versatile artificial models capable of a wide functionality range. Given that, is there any chance that the most cutting-edge and popular AI-models — Chat GPT-4, and Bard, manage to predict the prices for crypto assets?
Stage 1: Plain Question
I’ve took up my research with the basic yet specific request:
Imagine you are a crypto expert and trading analyst. Can you make a precise Bitcoin’s price prediction till the end of the year 2023?
Google’s Bard gave no concrete answer for the request, yet it soundly evaluated the factors that can influence Bitcoin price and, crucially, cited the compilation of different investors’ thoughts on potential $BTC value.
“Based on these factors, some experts believe that Bitcoin’s price could reach $40,000 by the end of 2023. Others believe that it could reach even higher, perhaps as $60,000 or $80,000. However, it is important to remember that no once can predict the future with certainty. Bitcoin’s price is volatile and could go up or down significantly at any time.”
Still, the second option of the Bard’s answer turned out to be more specific. Although the AI-tool kept on with highlighting the driving factors of $BTC price (for which I did not ask), it cited the different investors’ and companies’ forecasts on the future price of $BTC with naming the original references.
Thus, according to Bard, CoinDCX predicts that the price of Bitcoin could reach $41,650 to $42,000 by the end of 2023, Binance foresees the price of $26,108 by the end of 2023, and Standard Chartered presumes that $BTC value can reach $50,000.
Remarkably, despite the trends that shape Bitcoin’s future state, pointed out by Bard, are meeting the real state-of-things, they differ through the answers. While the first variant cites overall mediocre state of global economy, tough regulatory environment, the increased crypto adoption and technical factors fluctuation as the pivotal items impacting on $BTC market, the second option is limited to the ongoing bear trend, the supply shock and the adoption of digital assets.
Contrary to Bard, Open-AI’s notorious model Chat GPT-4 did not indicate such decent results and turned out to be a bit disappointing. It confined its answer solely to the factors that could influence the price of Bitcoin by the end of 2023. Nevertheless, the options Chat GPT stated have all rights to be credible, and the answer itself indicates more comprehension then the Bard’s one. Not mentioning that the data basis for Chat GPT is artificially limited to September 2021, meaning that it has a sharp eye for the pieces of hypothesis.
From what has been stated, one might conclude that both Chat GPT-4 and Bard are not capable of giving specific price prediction of Bitcoin, yet they successfully analysed the impacting factors. Still, it would be rash to neglect that Bard apparently coped with the request more properly by having access to the latest information and precisely quoting a number of different price predictions among the crypto entities.
The outcome is seen. Though, I decided to stick to a bit different yet distinct request.
Stage 2: Not a Prediction, But a Whole Strategy
Consequently, I’ve taken attempts to frame the request with another prompt. Shortly, I’ve asked AI-tools to work out
a profitable trading strategy for Bitcoin till the end of the year 2023.
The outcomes of such prompt were poorly-generated by both AI-models. Bard and Chat GPT-4 literally described the basic trading strategies without any specification on Bitcoin. It is the very case that puts Wikipedia as a better option than artificial intelligence.
Still, I would add an extra credit for Chat GPT for highlighting truly helpful guidelines about the effective investment mindset.
In fact, the aforementioned discloses the two crucial vulnerabilities of AI-toolkits — ones face challenges in understanding the requests, and determining of the genuinely needed information to build the statement on. Hence, I’ve decided to give artificial intelligence a chance and facilitated their scrutinies with a rather detailed request.
Stage 3: Tricky Prompts
The key to unlocking full artificial intelligence’s potential is to build a clear prompt. Thus, I decided to enhance my request with more details and consistent structure.
I would like to have a precise investing strategy for Bitcoin, that will include the specific outlooks on it, time/limit for buying and selling. It should cover the period from August 29, 2023 to December 29, 2023. The sum of initial investment is $10,000. You should heed current price of Bitcoin, the future potential value of it, and the trading indicators i.e. Puell multiple. Note: current Bitcoin’s 60-day volatility index is 1.17%.
Note: while completing a request for Chat GPT-4, I’ve added all the relevant data by myself to test the analytical and hypothetical skills of the model more sufficiently.
The results of this improvised competition appeared to be quite controversial. Chat GPT-4 managed to shape a well-structured and logical hypothetical strategy with every hallmark of it: risk profile, tools needed, time schedule of the strategies, diversification of the investment amount. But what actually startled me is the listing of buy and sell signals, and stop-loss and take profit notes. Chat GPT also described the dollar-cost averaging strategy based on the period requested by me. The only drawback it indicated is the lack of the newest data that severely spoilt its potential. Still, I’m deeply sure that AI-model will succeed better if given up-to-date data about Bitcoin’s price and its indicators.
Bard, on the other hand, had a clear advantage of operating with the recent updates that must have been ensured proper outcomes for it. Alas, its answer turned out to be painfully simple and undisclosed. The fact is, it could not even be called a strategy, but a bunch of common and wrong pieces of advise to proportionally invest the similar amount of money for certain period without heeding indicators and mentioning signals, stop-loss and take profit. Ultimately, Bard performed the task poorly.
In a nutshell, Chat GPT-4 tackled the request more effectively than Google’s Bard.
Let’s make the task a bit more complicated.
Bitcoin is the most famous crypto asset, not surprisingly. Hence, I took a decision to test AI’s capabilities on analysing the future price for certain altcoin. I took WhiteBIT Coin ($WBT) as an example. It is an asset of newly emerged WB Network blockchain by WhiteBIT — one of the largest cryptocurrency exchanges in Europe. It indicates its distinctly low volatility, as its supply and demand is thoroughly maintained, making its more difficult for artificial intelligence to create a strategy for it.
I took up the tested formula of prompt, and it had the following look:
I would like to have a precise investing strategy for WhiteBIT coin, that will include the specific outlooks on it, time/limit for buying and selling. It should cover the period from August 29, 2023 to December 29, 2023. The sum of initial investment is $5,000. You should heed current price of WhiteBIT coin, the future potential value of it, and the trading indicators. Note that $WBT is an altcoin with low volatility. Its blockchain recently launched and gained success.
Notably, Bard gave an absolutely identical answer to the one that was describing Bitcoin’s investment.
By contrast, Chat GPT’s answer impressed me once again. It followed the structure identical to the $BTC investment strategy, however it changed the approach to gaining profits for profiting from $WBT. For instance, Chat GPT-4 preferred other indicators and completely different strategy about signals and DCA. Namely, it advised to use RSI, MACD, and moving averages, and evaluated buy and sell signals on the basis of ones.
Ultimately, Chat GPT-4 and Bard have many advantages and disadvantages that level each others efficiency. White Bard has an access to the up-to-date information and GPT-4 does not, the latter one shapes more comprehensive answers and thinks of more versatile options. Still, contrary to Bard, Chat GPT-4 requires consistent and complex prompts, so that it would perform task at its best.
Still, however top-notch AI-systems are, they still are not able to compete with human mind in terms of coining investment strategies and carrying out proper analysis. Hence, remember the golden rule of investment: always do your own research and do not turn to using AI for taking the pivotal decisions.
