NEW GEMINI 3 AI Good For TradingView Strategies? (watch ASAP)

That's right, guys. Google launched

Gemini 3 and I've tested it for trading strategies and I've given it a very hard time. So, if you watch this video, you

time. So, if you watch this video, you know exactly what is good for and what is not good for. Can it convert an indicator into a trading strategy efficiently and with good results? And

spoiler alert, pretty good. Can it also improve an existing trading strategy to squeeze out better performance? Can it

create a new trading strategy from the scratch with good numbers? And the

newest challenge in my test suite is can it recreate a closed source indicator?

And the results are quite fascinating.

And in my opinion, you have to watch this video until the end. Also because

you will know how to use this new AI to your advantage, especially because everyone is saying is the best AI now.

So let's see if it is true. And if

you're a subscriber or returning viewer, then you might wonder where is Michael again? This is not the usual background.

again? This is not the usual background.

Well, yes, this is my office. So, I put out a studio in my office because the living room where I usually film is busy. Let's just say my wife is there.

busy. Let's just say my wife is there.

So, today this is the scene of the crime. And guys, make sure to like this

crime. And guys, make sure to like this video if it adds value to you because otherwise the YouTube algorithm will never know. And also, you're supporting

never know. And also, you're supporting this channel. Also, remember this is not

this channel. Also, remember this is not financial advice. Crypto is risky. You

financial advice. Crypto is risky. You

can lose all of your money, guys. And

that is why I run this educational channel so that you are always ahead of the herd. Subscribe and let's go. And

the herd. Subscribe and let's go. And

first of all, I asked it, how are you better than the previous Gemini version?

And on top of giving me a long text, it also gave me a table, which is very nice because it makes it very easy to compare, right? So you have Gemini 1.5

compare, right? So you have Gemini 1.5 and also Gemini 3, the current flagship model, and it says in reasoning it's PhD level. Amazing, right? deep think

level. Amazing, right? deep think

feature. Awesome. In coding, it says it's capable of vibe coding, which I don't really care to be honest. I just

want it to code very well and not make mistakes, but also build apps and complex features. Okay, let's see if

complex features. Okay, let's see if that helps us. Aentic, but that is basically just saying that it is able to create a complex plan of action before going at it, right? So, it's not really

agentic in the sense that it takes over your browser or anything like that. and

multimodality which could be super interesting when giving the AI multiple files to work with screenshots and CSV exports and other things. This could be

very interesting. And then I went ahead

very interesting. And then I went ahead to ask it, how are you better than Grock 4.1 and Cloud 4.1 Opus? Because of

course, right, those are very good AIs.

I know that already. And again, it gave me a nice table to make it easy to understand. And the main thing here is

understand. And the main thing here is that it says it has infinite context.

Well, that does not exist, right?

Nothing is infinite in computer science, right? But it says it has a million

right? But it says it has a million tokens, two 2 million tokens, that is if you pay for the pro version, right?

Which is way more than the 200,000 tokens include and 128 to 256,000 tokens in Gro 4.1. So, this should be very helpful. And again it tells us that it's

helpful. And again it tells us that it's good in multimodality working with different types of files text but video audio images as well which is awesome which can come in very

handy. And then it says best use case

handy. And then it says best use case complex analysis massive files which we have obviously we're going to export chart data and trades and everything and scientific

research. What I would have loved here

research. What I would have loved here is that it says it's good at coding right because that's what we need here.

But I'm happy that it says it's good at analysis because that's what we also need for it to run back tests and analyze its own creations, right? And by

the way, fun fact, it says that Claude is good at refactoring code and all but also at nuanced pros, which is a bit of a stab in my opinion, right? And then

when it comes to Grock, it's also a little bit funny because it says it has access to current events, of course, because it has the current tweets, but then also unfiltered chat. Yeah,

exactly. like Elon. So far it sounds amazing and also in the benchmarks and it's a lot of benchmarks here all combined into this result and you can

see it's in front of GPT 5.1 in front of Grog in front of Claude in front of everybody. Let's see if it helps us with

everybody. Let's see if it helps us with trading strategies. First of all I

trading strategies. First of all I wanted to convert this indicator behind me into a trading strategy and I wanted to decide fully by itself when to go in

and when to go out. Now this indicator is super interesting. I found it on Trading View and it is called Mystic Pulse V2.0.

Let me show you also about this script.

So this is the name of the indicator. I

will put the link down below. Why is

this indicator interesting? Because in

my opinion it has a very nice indication when it turns green that you should go in. And then as well it has a very nice

in. And then as well it has a very nice indication that when it falls off a cliff it turns red and then it's a good idea to go out. Right? So that was a profitable trade here. Not financial

advice just showing you the historic performance of this thing. But we don't yet exactly know the KPIs until we convert to a strategy. But it already looks interesting enough to do that.

Right? So here you can see it turns green. There is a buy signal here.

green. There is a buy signal here.

Right? And then it turns red over here.

This might have been like a very tiny profit, right? But then again, it turns

profit, right? But then again, it turns green here to buy back and then it falls off a cliff here and basically this is a sell signal. So that was another

sell signal. So that was another profitable trade. Let's see what Gemini

profitable trade. Let's see what Gemini 3 does with this indicator. Can it

produce a profitable trading strategy and is it better than the ones we already have access to? And for that I'm going to show you the prompt I was using which you also have access to. The link

is down below to this document. So it is called the convert indicator to strategy prompt and it basically says please convert this indicator to a trading strategy as you see fit. Cool. So I give

it all the freedom in the world. I also

tell it hey you have access to the chart data which of course I will give it the exported CSV file of the chart. I will

show you how to do that. And then I'm telling it look it's BTC 1day chart.

Okay. Avoid a couple of things and respect these things here. long only

trade from 2018 into the future one day chart 100% in and out that makes it easier to compare and then 0.1% commission one tick slippage and I paste the code of the indicator this is how it

looks like in Gemini as you can see this is the prompt and this is the code of the indicator and also I've attached the file the chart data which I will show you now how to do go back to trading

view and zoom out so that you have enough data from 2018 into the future at least because when you export the data going here by the way and clicking on

export chart data it will export only what you have loaded in the chart right so make sure to zoom out to have enough data loaded and also make sure to turn off all the other indicators that you

have even this one because otherwise it will be part of the export for the CSV it has too much data in there that confuses the AI okay so now because I have turned everything off I have only

the chart in front of me this is Bitcoin one day chart index And so export the chart data and then export. And when you open up that CSV file, it should look something like this, right? It has time,

it has the open, high, low, close, prices of the day. And it has no other columns on the right side. It would have right more columns if you have turned on volume or you have turned on this, but

it would have too much data and the AI would get confused which data to use. So

I always make sure that I export only the pure data of the asset and nothing else. Okay. Okay, now let's copy the

else. Okay. Okay, now let's copy the prompt and go to Gemini. We open a completely new chat and we paste it in there. And by the way, we make sure that

there. And by the way, we make sure that we have the latest model selected. It's

the thinking with three pro. That is the default. Just making sure. And now below

default. Just making sure. And now below this, you have to paste the code of the indicator that it should convert, right?

So go back to trading view, hover over the name of the indicator, and then click on the source code brackets. And

as you can see, there is a panel opening up on the right side. Just copy the entire code. Go back to Gemini, paste it

entire code. Go back to Gemini, paste it here, and let's go. And I'm not going to let you wait until this thing finishes because I already have a result. Okay.

So, it gave me the source code, but it also gave me a diagram showing me when it buys, when it sells, the equity line, and the indicator itself, and like it's very interesting. And then it also gave

very interesting. And then it also gave me actually here it gave me the total return. It did a back test on its own.

return. It did a back test on its own.

Let's see if these numbers match at all.

And then it gave me this totally useless graphic, which I will ignore, but uh it's very strange. And it gave me some explanations. And it gave me the code.

explanations. And it gave me the code.

So, let's copy this code and let's go to Trading View. And because now the entire

Trading View. And because now the entire UI of Trading View is a bit different.

Basically, what you have to do is you can click here on Pine Editor, right?

and it opens up on the right side or you can click this icon right here which opens up the same thing. And then here you can click to create a new strategy

and then you can just paste it in here.

Click save and then click save again and then add to chart. Tada. We have KPIs for this strategy. But before we look at the KPIs actually let's see if it is

trading the way we would be trading.

Okay? because we said that we wanted to buy when it turns green and sell when there is a cliff. And as you can see, it does exactly that, right? So this is the buy signal, the arrow pointing up. And

this is exactly where it buys. And then

here where the cliff is, exactly after the cliff, it sells exactly here at the candle open of the next candle, which is totally normal. It also has some losing

totally normal. It also has some losing trades on sideways markets, but that is normal, I would say. So this is pretty much exactly what I would have expected.

Now, let's check one more trade. It goes

in here when it turns green and it sells after the cliff. That is pretty nice.

And before I show you the KPIs, I want to say to YouTube, this is not an indication of future success. I'm just

showing this nice people the KPIs of this nice trading strategy. Okay? So,

not financial advice, guys. All right? 3

2 1. Those are the KPIs. Let's go. This

is a profitable strategy now. Okay?

2,497% profit with 46.12% max draw down. The max draw down is a bit too much for me because on Bitcoin 1

day chart it should be below 30%. But

okay fine, right? So the first trade if I scroll way down here on the list of trades you can see was on January 6, 2018. Okay, so since then until today

2018. Okay, so since then until today which is the time of this recording these are the KPIs. Now that is not the important part. The important part is

important part. The important part is how does it actually perform on altcoins and you'll be very surprised. This is

how it performs on Ethereum 1day chart and it's crazy good. Okay, 9,300%

profit with 45% max draw down. Again,

the draw down is a bit too much but mate the profit is very good. Not financial

advice, not promising you the same thing in the future, but very exciting. Let's

check out Solana as well. on Solana

guys. 58,000%

profit with 66% max draw down. Again,

the draw down is high. Profit is

massive. I mean, it is crazy good. Let's

go to the next one. Cardano. Okay, on

Cardano, it's not that great because usually strategies here make 5,000 6,000% profit. So, 34% not the best. My

6,000% profit. So, 34% not the best. My

strategy made 5,939% profit with 36.16% max draw down. For example, what about

Suie? 97% profit with 53% max draw down.

Suie? 97% profit with 53% max draw down.

Also not the best. For example, my strategy made 495% profit with 27.88% max draw down in the same time frame.

But I'm still quite impressed with the results it had on Ethereum and Solana.

And I'm sure that if you use it on your portfolio coins, you might find nice combinations. And the best part is I

combinations. And the best part is I will give you the code of this strategy.

It will be also in the document down below. The next challenge is to take an

below. The next challenge is to take an existing strategy and improve it. And in

this particular case, you all know the goshion channel strategy. It's quite

famous meanwhile because it works quite well. Not financial advice, but I'm

well. Not financial advice, but I'm talking about this strategy for at least eight months now on my channel. Everyone

has access to this strategy and I can even show you how to get access later on. But the thing is that this strategy

on. But the thing is that this strategy goes long only. It's quite a good strategy. But a lot of people were

strategy. But a lot of people were asking, hey, can we also go short because there might be opportunities here to make some additional money, right, by going short the market. Now,

so far, whenever I was experimenting around with this idea, I never really found a good way or a good logic that adds value because if it doesn't add value, why would I go short, right? So,

let's ask this Gemini 3 AI to add shorting logic to the strategy to make it better. And I'm going to go through

it better. And I'm going to go through this a bit faster because we have a lot to go through. Okay, so scroll down in this document. The link is down below.

this document. The link is down below.

So this is the prompt for improving a strategy to add shorting logic. Okay,

basically it says please improve the strategy by adding logic to also take short trades to benefit from falling prices. So on and so on. And here is the

prices. So on and so on. And here is the additional thing. I add the chart data

additional thing. I add the chart data but I also give it the list of trades of the existing strategy so that it can see how it was performing in the past so

that it can actually improve it. My idea

here is that it will run a back test and then it has the data from the original strategy and it can compare with what it has produced. In fact, I'm even telling

has produced. In fact, I'm even telling it to run a back test internally and give me the best version you can come up with. So I have to attach three things.

with. So I have to attach three things.

the strategy code, the strategy trades, and also the chart data. The chart data you already know how to get. But to get the strategy trades, you simply open the

strategy tester, and you download the data from here. Click this download data as Excel file. And this is how it looks like. It gives you the entire

like. It gives you the entire performance, the trade analytics, and all these things. And what I care about is the list of trades, which is one of these tabs. And I'm not going to mess

these tabs. And I'm not going to mess with this file whatsoever. I just want to make sure that the data is there. And

this is the file I'm going to give to the AI. I'm going to go back to the

the AI. I'm going to go back to the prompt. I'm going to copy the whole

prompt. I'm going to copy the whole thing and I'm going to go to the AI and give it the entire prompt. And then of course, I'm going to paste the code of the strategy I want to improve. Now, I

have the code of the gshian channel strategy which I put here. You don't

have it. It's fine. I'm just going to show you what the result is. And of

course, I've given it the chart data and also the trades file that I showed you before. So, scrolling way down, I got a

before. So, scrolling way down, I got a strategy from the first try. So, we're

going to copy this strategy code and see how it performs. Back in Trading View, these are the original KPIs of the Goshion channel strategy without any

shorting logic. Remember, 2,700% profit

shorting logic. Remember, 2,700% profit with 30% max draw down. Okay, so now let's go to the Pine editor. I click

this icon over here and I have a test strategy where I always paste the latest AI code in there. I save it and I also have to click add to chart and once I do

that I get the numbers of that strategy.

Well, it looks pretty bad guys. All

right, because 150% profit with 37% max throw down. H I think it's pretty bad.

throw down. H I think it's pretty bad.

And if you come here, right, and you can see the performance. Let me put it way up here so I'm not in the way. You can

see that the shorting logic actually loses money while the long logic makes money. Not a lot. Yeah, but actually the

money. Not a lot. Yeah, but actually the shorting logic is not beneficial whatsoever. And of course, I'm not

whatsoever. And of course, I'm not impressed and I got a little bit annoyed, right? So I went to other AIS

annoyed, right? So I went to other AIS like Grock, Claude, and Chet 55 to check if they can produce better results when it comes to shorting. And I have all the

results from the different AIs right here, guys. So this one is Gemini 3.

here, guys. So this one is Gemini 3.

Okay, these are the numbers you have just seen. And this is JP5. You can see

just seen. And this is JP5. You can see it's able to produce a little bit better numbers, more profit, but also higher draw down. Not very impressive. The next

draw down. Not very impressive. The next

one is Grock. And Grock is able to produce a little bit better numbers and also the draw down is lower than the CHP one. So that is fine. But the best

one. So that is fine. But the best result was produced by Claude and it made a profit of,35%

with 47% max draw down. But still not impressive because without any shorting logic, this strategy is way better off and also takes less risk. At this point

in time, the AI cannot come up with a good shorting logic for this strategy. I

cannot do that as well. So, it is what it is. Let's go to the next section and

it is. Let's go to the next section and figure out if the AI Gemini 3 is able to come up with the trading strategy from the scratch. And this is going to

the scratch. And this is going to unfortunately be a quick one because this AI fails at this task. Okay, so

this is the prompt. By the way, you have the prompt also in the document linked down below. I need the best trading

down below. I need the best trading strategy you can come up with for Bitcoin one day chart. The data is attached. And of course, I attach the

attached. And of course, I attach the data, right? So this is the data. This

data, right? So this is the data. This

is the prompt. And whenever I let it run, okay, it never finishes calculating. Actually, it crashes after

calculating. Actually, it crashes after like a few minutes. It comes back to the prompt being down there. and nothing

happened and even if I reload nothing happened. So the whole thing doesn't

happened. So the whole thing doesn't work. So this is a fail and we can go to

work. So this is a fail and we can go to the next task which is to reconstruct an indicator that is closed source and to basically recreate it. So you have the

code of a closed source indicator. It's

crazy stuff, right? So it works and I already have a video about that. I will

link it up top. It's quite popular. But

let's see if this AI can do it even better.

What is happening with this AI, guys?

Look what it tells me. I'm having a hard time fulfilling your request. Can I help you with something else instead? No.

Okay. I want you to recreate this indicator in Pinescript. And I know it worked already with Claude because I showed you that in another video. I told

you it's up there on the right side. So

unfortunately I think that Google doesn't give this new AI enough time to actually finish the job where Claude can work for half an hour even and come up

with something. This one is get cut off

with something. This one is get cut off and that is why I'm getting this failed result over here. So unfortunately this is another fail for Gemini 3. So guys

what's the conclusion? Is Gemini 3 any good for Pinescript trading strategies?

And I would say it's totally useless.

Okay? Because I have to tell you honestly, it was only able to manage one challenge properly, which is the one to convert an indicator into a strategy.

But then again, JPT could do that like 6 months ago or even more. I have a video up top, 1.4 million views, converting indicator to strategy. But now really any AI can do it, right? Like so if I

have to solve complex problems when it comes to trading strategies in pine script for trading view I go to grock and I go to claude and that's it. Okay

any other AI will have some sort of issues and claude is very powerful because it's allowed to run for a long time. So when the task is heavy no

time. So when the task is heavy no problem it will continue running for a long time. Yes, it will consume tokens,

long time. Yes, it will consume tokens, but you can buy more tokens so you can solve the problem, right? Instead of

getting cut off or the AI telling you that, sorry, I cannot fulfill your request. I mean, what is this? If your

request. I mean, what is this? If your

crypto portfolio is $25,000 or more, then you need more than just trading strategies. Consider the autotrading

strategies. Consider the autotrading masterass where I teach you my entire process, including my trading strategies, of course, which are the best in the industry. Click on these screenshots and you will see every

single trade they took. But you need the education to properly build a riskadjusted portfolio to know how to play the markets depending on the cycles and how to rebalance, how to use all of

this to your advantage. Right? Do your

homework by the way. Check the reviews, watch the video here on the website and you will understand that this is a limited access program for thousand people only. We still have some seats

people only. We still have some seats because I just started the program one year ago. This is the most advanced

year ago. This is the most advanced program that you can find. And this is just for higher net worth individuals, for people who have enough money to understand that risk management is the

most important thing. Okay? So, if this is for you, apply for access. Click this

button. We're going to talk to each other. If it's a good match, then you

other. If it's a good match, then you can join. That's all I have for you guys

can join. That's all I have for you guys in this video. And if you want to learn how to create a trading strategy with AI from A to Zed, then this is the next video you should watch. It's trade and

[music] view.

Uh-huh.

Oh yeah. Oh yeah.

>> Strategies. [music]

We know what to do.

Yeah. Do you [music] stop losses set? Watch the bus rise and