¶1So about 5 days ago, I created my building a GPT6 cali AI trading broth from scratch. My full guide that I put up on YouTube. So today I thought we just look in and see how this has been doing. So if you didn't watch the video, you can of course go back and watch it after this. Uh but basically this is my Cali AI trading bot that is using the new Google weather next [music] tree from Google Deep Mind to try to predict some weather markets.
¶2And in the video, we only focus on the New York City weather. And we did create like a VPS system that runs 24/7. I had some few issues when I was out traveling and I had to do some adjustments, but other than that, it's basically been running every day. And I thought we can take a look at some of the results today and catch the market that is going live in about 1 hour. I think little less than 1 hour.
¶3So, why don't we just take a look at how we have been doing so far and we're not going to spend too much time on this video, but I just wanted to give you a quick update if you did follow this or similar guide to see how I have been doing. So, yeah, let me pull up how we have been doing so far. Okay, so here you can kind of see how we have been doing. So, September 10, I think that was the first day I ran it, I think. So, uh we actually lost some money.
¶4I there was a bit of an issue here, but basically that position lost. I think it should have been a bit more than $2 down, but uh we got exited something and yeah, something happened. At least we lost the first day. Uh September 11th, that was Friday probably. [snorts] We uh let's look at this.
¶5We had uh what did we have? Yeah, we had 79 to 80. You can see + 12. uh plus 12. It's It says plus 10 here, but this is after fees and stuff, I think.
¶6So, we did uh on 11th of September. Oh, that's in New York City. That's a bit of a historical date. Uh $12 plus. Um this is pending.
¶7This is from yesterday. But, uh you can also see here uh plus 9 uh 77 or below. So, this also landed. Uh but the most interesting is actually the prices we got. We might take a look at that afterwards.
¶8Uh this is from yesterday I think or like today or something. So we entered at no uh but this is basically in because it's already past 73. So this is a 99. If you take a look at this my position here, right? You can see we entered at 57.
¶9So that's pretty good. I think 57 here is really good and unrealized is 1403 [snorts] but this has not settled yet but uh it should be it is basically settled. So we made like $14 on that and yeah that's basically how far we have run this so far. So if you're going to count all the markets that has not not been realized [snorts] we are like $34 in plus. So yeah pretty good start to this.
¶10I'm going to keep running it and I might expand to other markets, but for now I'm just going to keep it to New York. And here you can kind of see yeah all the contracts and stuff like that. So basically that today uh I wanted to do just um go through everything again. So I kind of prepared for September 15. This is tomorrow.
¶11So we're going to prepare to for today. We have armed the model and yeah we armed the model live. It's live now on the VPS. And this is a bit more interesting because this bucket is really close. So you can see we have 71 to 7 to 71, 72 to 73.
¶12These are really close. And it's going to be interesting to see what happens here. So I'm just going to ask our GPT6 here. So what will our system most likely enter here? Uh basically I think I set it up so we we need some kind of plus expected value.
¶13And I haven't seen exactly this setup yet. Right. So, it's going to be really interesting to see what bucket we land on here. If we just go for um probability, we are probably going to land uh it's kind of hard to say. Maybe in this one.
¶14Let's see. So, you can see here there's no pre-selected bucket. The bot will choose the side with the largest conservative edge at launch because it doesn't really matter too much about the prediction here. Uh as long as we don't have the prices, right? The likely candidates are yes uh on a middle bucket or no on a hotter bucket.
¶15But launch decide launched price but launch prices decide. Yeah, that's was what I was going to say. So everything is about the price, right? Uh so we want to pick where we have most conservative edge and the minimum is 5 cent. So I did explain that in the previous video.
¶16So let's say conservative yes here is 34 cents. Uh that's that's our model right that means that um the price on cali must be like 28 right we need to subtract five here it's going to be around 34 and stuff like that. So, this is going to be maybe some kind of long shot. We'll just have to wait and see when the price is open. So, that is what we're going to do today.
¶17Uh, I think we are 45 minutes away. So, I think we're just going to wait for that. And yeah, maybe I'll show you one other thing in the meantime. This is what I wanted to show you. So, these are kind of the entry edge if you want to call it that from our model.
¶18So, it looks really good, right? Uh, it kind of says on this entry we had 16 cent of edge. Here we have only seven. That was really low. 27 and 22.
¶19I think that's a bit too much on average. 18. It might be a bit beginner's luck here. And this will probably flatten out over time or we have just gotten lucky. But we can't really tell much after only five days.
¶20So I would probably come back in like I I guess we need like 30 days at least just to see how we have been doing. So I might actually try that. I might write try to run this for a month and then we could do a new update if people want that. Give this video a like. Subscribe if you want to see an update on my um Kali New York City trading bot in like 30 days.
¶21Could be interesting with this great start. Let's just see if it can continue. I expect some draw downs and stuff like that but uh yeah could be interesting and the entry edge has so far been doing pretty good I say I would say uh we can also check the prices here so you can see we enter at 56 66 47 and 77 so yeah 10 minutes no 10 minutes 30 minutes until we go again so yeah see you then okay so now we are just like 4 minutes away So I asked what predictions did we freeze the model on? So like 15 minutes before open we actually freeze the model. So you can see it kind of changed a bit at the end.
¶22Now we have the 72 to 73 bucket has a conservative edge of 42. So this is the strongest prediction. So now we just going to have to check what buckets come up and what the prices are. So if any bucket has the best conservative edge. So let's say no uh on this one has more than yes here.
¶23Let's say this has like 10 cent edge. This has eight. Then we're going to do no here. Oh yeah, it's probably not going to be here. But uh yeah, it could be something like that.
¶24We're just going to have to wait and see. And yeah, that should be like 3 minutes away. Uh if we don't find 5-cent edge, we just going to pass and wait for tomorrow. So hopefully just for the video sake, we're going to find something. But yeah, we know that in 4 minutes.
¶25Okay, so it's now uh 1501. Uh let's just refresh here and see if we did enter on any position here. Uh we may not do that. Uh let's just wait and see a bit. I'm going to ask here.
¶26So just give me the URL to the event. No, not yet. So this is the event. So let's see here now. So you can see the bot made no trade at 1501.
¶27Uh it ran successfully but no offer made our 5cent edge. So that's just the way it is. And you can see we were very close on no below 69 at 85. So that was very close but that is the point about having this system. Uh I also said check the pricing now because now u the edge has even moved further away.
¶28Right? You can see that. So, but that is basically of course happened today, but basically this is the point of having this automated system. Some days you find an edge uh that kind of fits with your model and the prices on the event. Some days that there is no edge and then the best expected value is to do nothing.
¶29just wait for the next day and then we are going to avoid very unsecure uh yeah positions like maybe today we don't know but uh that is basically I'm you should be as happy with making no trades no qualifying trades as making a good one I think so yeah that was just the way everything rolled today so a bit unfortunate for the video sake but uh yeah tomorrow there's always a new chance I guess so just to sum this up. Um, the bot is working. It's running super smooth on the VPS. So, everything works. That's the most important part.
¶30And it's also good to see, like I said, that we don't make trades when there is no would you say edge edge according to our model. You can say that. Is there a real edge? We don't know yet. So, according to our model today, there was no edge and we didn't do any bets.
¶31So, basically, uh, I might have to just come back in Yeah, we don't have to do 30 days. We can come back in like 14 days just to check in if I keep this running. It could be interesting. Just do a short video, check the results, maybe try to run it a bit more if everything is running well. So, if you want to know how I set this up, just go back to the previous full guide video and you can try this out for yourself.
¶32Uh, if you want to support the channel, maybe follow the link in the description to Cali and check out my other links if you want to. So yeah, just a short update on how the bot has been doing the first week or something. I guess about a week. So yeah, thank you for tuning in and thank you for the support. And if you like this content, give it a video a like and subscribe.
¶33And I'll see you again soon.