AI - BTC (2) The second branch done in BTC is to use aggtrade data to create a local replica of exochart. The ultimate goal is to create an automated monitoring system for the entry model of exhaustion reversal in my subjective trading .. In the daily market of the past week, I also posted a similar interface to Figure 1, which is the interface of a trend reversal line similar to exochart that I ran locally. The purpose is to capture various volume delta reversal events to assist in the judgment of entry models. From the current results, the signal quality is quite good. Figure 1 is the most recent example But in terms of signal generation I have also developed several systems . There is a version with higher frequency but lower signal quality, which can capture many reversals, but not every reversal is a true reversal as shown in Figure 2 There are also versions that only filter high-quality reversals for inclusion in Figure 3, but this may not yield a signal for several days .. There are also balanced versions . It depends on how much time each person has available for staring at the disk . The current signal system still has a long way to go before automation, as this is just an entry model. After entering, it depends on the market and the design of SL TP So the subsequent automation design after entering the model This is the next step of research .. Whether the current filtering can be effective in the long term or overfitting in the short term also needs further verification . There was originally a very good version with 6-7 factors, but the filtering conditions were basically 100% overfitting. I gave up directly. The current versions only contain a small number of conditions. Now everything is more on Discord . The same idea has been applied to ETH now, and we plan to start testing next week. If I have time next week, I will also do the gold version mentioned earlier. After all, there is tick data from COMEX GC. Doing this is all about reusing existing frameworks. Not too difficult.