I Turned My Trading Strategy Into a Bot — Transcript

Full transcript

1. 0:00I ran an experiment. I took one of my
2. 0:02trading strategies, turned it into a
3. 0:04TradingView indicator, made sure it was
4. 0:06finding the exact trades that I wanted
5. 0:09to. So, it had a one-to-one visual
6. 0:11representation of what I see when I look
7. 0:13at the chart and what the indicator is
8. 0:15telling me. And then I took the exact
9. 0:17same rules and turned them into a
10. 0:19completely automated trading bot on
11. 0:21MetaTrader. Same strategy, same entries,
12. 0:24same stop losses, same take profits. At
13. 0:28least that's what I thought. Because
14. 0:30when I started testing the bot,
15. 0:31something really strange happened. It
16. 0:33worked. It was taking trades. The
17. 0:36backtest was profitable. Some of the
18. 0:38numbers actually looked really good. But
19. 0:40when I started going through the
20. 0:41individual trades, I realized the bot
21. 0:44wasn't always trading the strategy I had
22. 0:46built. And one of the mistakes it made
23. 0:48was so ridiculous that it completely
24. 0:50changed the way I think about automating
25. 0:52trading strategies. Now, let me show you
26. 0:55what happened. The strategy itself isn't
27. 0:56really the important aspect of this
28. 0:58video, but I'll explain it quickly so
29. 1:00you can understand what the computer was
30. 1:02trying to do. I made a previous video
31. 1:04about this, and I've traded RSI
32. 1:05divergences for years. One particular
33. 1:08setup that I like combines hidden RSI
34. 1:10divergences with a Fibonacci
35. 1:12retracement. Basically, there's a move
36. 1:13in the market, price pulls back, and I'm
37. 1:16looking for a very specific relationship
38. 1:18between price, RSI, and the 0.618
39. 1:22Fibonacci level. When everything lines
40. 1:24up, that's my trade. Now, normally when
41. 1:25I'm trading this myself, it's incredibly
42. 1:28visual. I can look at the chart and say,
43. 1:30"That's the move. That's the high,
44. 1:31that's the low, that's the retracement,
45. 1:33and that's the divergence." I look at
46. 1:35how the candle looks at the Fibonacci
47. 1:37level, and that's my entry. And that's
48. 1:39important because before I automate
49. 1:41anything, I first built it as an
50. 1:43indicator on TradingView. And if you're
51. 1:45thinking about using AI to build your
52. 1:47own trading system, I really think that
53. 1:48you should do this first. Don't
54. 1:50immediately ask AI to build you a
55. 1:52trading bot. Ask it to build you an
56. 1:54indicator that shows what it thinks your
57. 1:56strategy is. Because what you tell it
58. 1:58and how it interprets it two completely
59. 2:01different things. Now, the computer has
60. 2:03to show its work. This was basically my
61. 2:05first test. I explained my strategy and
62. 2:07had it plot what I see visually on the
63. 2:10chart. Whenever the computer thought
64. 2:12there was a setup, it showed it to me.
65. 2:14And I could sit there looking at the
66. 2:15chart and compare what the computer saw
67. 2:17against what I saw. Sometimes I'd look
68. 2:19at a signal and think, "Yep, that's
69. 2:21perfect. That's exactly what I would
70. 2:23have traded." Then I find another one
71. 2:24and think, "No, no, that's
72. 2:26That's not right." So, I'd figure out
73. 2:28why it was showing me that, change the
74. 2:30rules, run it again, and slowly the
75. 2:31computer started seeing the chart more
76. 2:33like I did. That's one of the reasons I
77. 2:35find AI to be extremely, extremely,
78. 2:38incredibly useful. I'm not necessarily
79. 2:41asking the AI find me a profitable
80. 2:42strategy. I'm saying, "Here's what I
81. 2:44already do. Help me explain it precisely
82. 2:47enough so that a computer can do it for
83. 2:49me." Those are two very different
84. 2:50things. Eventually, I got the
85. 2:52TradingView indicator pretty damn close
86. 2:54to what I see. So, now came the actual
87. 2:56experiment. I wanted to see what would
88. 2:57happen if I took those rules and gave
89. 2:59them to MT5, because now I could turn it
90. 3:01into an expert advisor, an EA, and test
91. 3:04the strategy mechanically longer across
92. 3:07more years of market data. So, I
93. 3:09translated everything over, and
94. 3:11initially, it looked like it worked. MT5
95. 3:14started taking trades. I had entries,
96. 3:16stops, targets, winners, losers, an
97. 3:18equity curve, all the stuff you normally
98. 3:20look at when backtesting a strategy.
99. 3:22But, there was one massive difference. I
100. 3:24couldn't necessarily see its thought
101. 3:26process nearly as clearly anymore. It
102. 3:29just showed me entries. On TradingView,
103. 3:31when I saw the trade, I could see why
104. 3:32the trade existed. It had the start of
105. 3:34the impulse move, the end of the impulse
106. 3:36move, the Fibonacci retracement level,
107. 3:38the level where it broke the previous
108. 3:40structure, the take profit, and the stop
109. 3:42loss, all on a single indicator. On MT5,
110. 3:44I basically saw sell, and then it
111. 3:46trades. And that's a problem, because I
112. 3:48started comparing the individual trades
113. 3:50between what TradingView was giving me
114. 3:52and what MT5 was giving me. And some of
115. 3:54them matched almost perfectly. Same
116. 3:56setup, almost identical entries, same
117. 3:58stop, same target, beautiful. But then,
118. 4:01TradingView would take a trade and MT5
119. 4:03wouldn't. Or MT5 would take a trade and
120. 4:06TradingView didn't. And I remember I'd
121. 4:08given the exact same strategy. So, how
122. 4:10is this possible? It turned out one of
123. 4:12the problems was something ridiculously
124. 4:14small. The swing highs and the swing
125. 4:16lows. As a human, this seems obvious.
126. 4:18You look at a chart and there's a swing
127. 4:20high. But the computer doesn't see a
128. 4:22chart like you do. Imagine two candles
129. 4:24next to each other with the exact same
130. 4:26high point. Which one counts? The first
131. 4:27candle? The second one? Both? Neither?
132. 4:30That tiny decision matters. Because if
133. 4:33the computer chooses a different high,
134. 4:35now it's comparing RSI to a different
135. 4:36point. That can change completely where
136. 4:38the divergence exists. That can change
137. 4:41where the impulse is measured. And that
138. 4:43changes the Fibonacci level. One tiny
139. 4:45difference at the beginning can produce
140. 4:47an entirely different trade at the end.
141. 4:49And that's when I started realizing how
142. 4:50dangerous this was all becoming. One
143. 4:52tiny little fraction decimal point of an
144. 4:54error blows the strategy. And the issue
145. 4:57is I wouldn't necessarily know it's
146. 4:58happening. Because if I had just looked
147. 5:00at the MT5 results, I wouldn't
148. 5:02necessarily know any of this was
149. 5:04happening. And this wasn't even the
150. 5:05biggest problem. My first version
151. 5:07entered trades at a market execution
152. 5:09once the setup was confirmed. And then I
153. 5:11tested that version for like eight
154. 5:13years. It had about a 75% win rate. And
155. 5:16now imagine that's where I stopped.
156. 5:18Imagine I posted the screenshot online,
157. 5:20"Trading bot gives 75% win rate. Looks
158. 5:23incredible." Except I started inspecting
159. 5:26the actual trades. The winners were much
160. 5:28smaller than the losers. Nothing lined
161. 5:31up to the original TradingView indicator
162. 5:34where it had perfect structure, perfect
163. 5:36placement for the stop loss, perfect
164. 5:39placement for the take profit.
165. 5:41Everything was ideal. And this was
166. 5:43taking the same concept but completely
167. 5:47skewing the structure of where the
168. 5:49impulse move was measured to and from.
169. 5:51And this is why I wanted to run this
170. 5:53experiment and make this video because a
171. 5:56lot of you guys are new into trading.
172. 5:59You are interested in making a
173. 6:01TradingView indicator. You want to
174. 6:04create the perfect setup for your daily
175. 6:06trading. And the issue is you must first
176. 6:09learn that specific setup so that you
177. 6:12can visually see it on the chart. Then,
178. 6:15once you have a concrete idea of what it
179. 6:17looks like, do your best to translate it
180. 6:19into something a computer can understand
181. 6:22and have it make version 1, version 2,
182. 6:24version 15 until it matches exactly what
183. 6:28you see and it shows its work on the
184. 6:30indicator. It shows you with the exact
185. 6:32levels what your structure is, what the
186. 6:35price movement is, how the RSI
187. 6:37divergences are working, what's bullish
188. 6:39structure look like. All of those things
189. 6:42need to be visually represented on the
190. 6:44chart. That's why I always recommend
191. 6:47start with a TradingView indicator
192. 6:49because a lot of you guys are going to
193. 6:50be tempted to immediately go into an MT5
194. 6:53EA, automate everything, walk away. That
195. 6:57is going to ruin your life entirely and
196. 6:59blow account after account after account
197. 7:02within the next month or so because this
198. 7:04is extremely important to me, I'm going
199. 7:06to do a tutorial video that's like an
200. 7:08hour long, my exact process of how I'm
201. 7:12going about structuring this and how I'm
202. 7:15prompting AI to get what I'm thinking
203. 7:18into its brain. But for now, what I want
204. 7:21you guys to learn, whatever strategy
205. 7:23you're working on or if you haven't
206. 7:24learned one yet, learn one. Then, master
207. 7:28it. Absolutely master it. Get something
208. 7:30that makes sense to you so much so that
209. 7:33you could explain it to a five-year-old.
210. 7:35Then, make an extremely wordy version of
211. 7:38that explanation to the five-year-old
212. 7:41and start talking to your AI. Whether
213. 7:43you're using chat GPT, whether you're
214. 7:45using Grok, it doesn't matter. What I
215. 7:48want you to do is use TradingView, use
216. 7:51AI, and use your knowledge in a cohesive
217. 7:54environment where you're getting almost
218. 7:56a one-to-one depiction on your
219. 7:58TradingView chart of what your eyes
220. 8:00visually see. Let me know in the comment
221. 8:02section if you actually want to see that
222. 8:04hour-long tutorial. If this video was
223. 8:07helpful for you, please share it with
224. 8:08your friends. Thank you so much for
225. 8:09watching, and we will see you in the
226. 8:11next one.