Crypto Coin Screener in Python — Transcript

Full transcript

1. 0:00Hey guys, so in this video I'm going to
2. 0:02share with you the code that I wrote for
3. 0:04a cryptocoin and time frame scanner in
4. 0:06Python. Now when will you need this?
5. 0:07Well, a good example of it is the
6. 0:09previous video that I did. It was about
7. 0:11this mean diversion strategy in Python
8. 0:13which was producing pretty good results
9. 0:14but only if you were trading it on three
10. 0:16different assets. But why just those
11. 0:18three? Why not other assets? Did I look
12. 0:20for them? No. And was even the time
13. 0:22frame that we were trading it on the
14. 0:24best option? Well, to answer this
15. 0:25question, I wrote this script to be able
16. 0:27to go through a lot of coins and time
17. 0:29frames very quickly. So, if that sounds
18. 0:32interesting and useful to you, let's get
19. 0:33right into it. Now, before I continue, I
20. 0:35got to say I am not a financial advisor.
21. 0:37This video is merely a tutorial and you
22. 0:39should always do your own research
23. 0:41before trading. And don't take this
24. 0:42video or any of my previous work as a
25. 0:44plugandplay money printing machine. Such
26. 0:46a thing doesn't exist and trading always
27. 0:48carries risk and you should only trade
28. 0:50if you accept full responsibility for
29. 0:52it.
30. 0:53So here's the result for the strategy
31. 0:55that I showed you in the previous video.
32. 0:57It was called mean reversion RSI and by
33. 0:59the end of it we were trading it
34. 1:01simultaneously on ETH BTC and Doge on
35. 1:04the hourly time frame. This was the P&L.
36. 1:06This was the maxonus 22% which is really
37. 1:09good. The Vet was 92% and the sharp was
38. 1:12almost two. So I'm pretty proud of this.
39. 1:14But the real question was why not other
40. 1:17coins, right? And was evenly the best
41. 1:20choice for all different coins. So to
42. 1:22answer this question, I went ahead and
43. 1:25wrote two scripts. The first one is for
44. 1:28importing data for so many coins
45. 1:30simultaneously. So this is it is called
46. 1:32batch importer. Feel free to name it
47. 1:34whatever. And by the way, I will share
48. 1:36the source code for both of these
49. 1:37scripts and the link will be in
50. 1:39description. So the first one is
51. 1:40responsible for importing data. And
52. 1:42inside it, there's this config section
53. 1:44which is almost the only thing that you
54. 1:46guys should care about. First give it
55. 1:48the name of the exchange. So this is an
56. 1:49enome to make sure you are giving the
57. 1:52right name. So for example, if I want to
58. 1:53give the name of Binance, you have these
59. 1:55options, Binance spot, Binance use spot
60. 1:58and binance perial futures and it's
61. 2:00test. So this is the one that I want.
62. 2:02But if you want something else such as
63. 2:03gate, you can also do this and you won't
64. 2:06have to worry about the exact spelling
65. 2:08of the name of it.
66. 2:10And next we have the list of the
67. 2:12symbols. And as you can see, I've given
68. 2:15it so many here, right? So just add as
69. 2:19many as you like. And next you should
70. 2:21give it the start date. Now a note on
71. 2:23this, it doesn't exactly matter. So just
72. 2:26give it a very early date. For example,
73. 2:28I've given a 2018 and many of these
74. 2:31coins didn't even exist back then. But
75. 2:34Jesse is smart enough to know that
76. 2:36sometimes it needs to import from after
77. 2:39the date you're giving it if there's no
78. 2:41date of before that date. So just give
79. 2:42it an early date such as 2018 and it
80. 2:45should be good to go. The other good
81. 2:47thing about the script is that it will
82. 2:49import them one after another. So it's
83. 2:51careful about the API rate limits of the
84. 2:54exchange and even if there is any
85. 2:56network issues for example it will keep
86. 2:58retrying and it will also skip the
87. 3:00existing candles. So you don't have to
88. 3:02worry that for example one of the coins
89. 3:05already exists on your database and
90. 3:08lastly when it is done with every coin
91. 3:11it will sleep for 24 hours and then try
92. 3:14to import the data for that single day.
93. 3:16So what I mean is you can just run this
94. 3:18script on your machine and let it run
95. 3:20for weeks or months and never have to
96. 3:22worry about importing candles again. So
97. 3:24if you are a bit lazy with you know
98. 3:26updating your database this is a pretty
99. 3:28good option. All right. So assuming that
100. 3:30we have the data now let's move on to
101. 3:32the actual script that is responsible
102. 3:34for running the back test. So again we
103. 3:37have a config section. All right. And
104. 3:39this is the only thing that you guys
105. 3:41should care about not every function
106. 3:42that is in here. And firstly we will
107. 3:45give it the name of the strategy. So in
108. 3:47the previous video it was called mean
109. 3:48version RSI. Next we have the name of
110. 3:50the exchange again. And then we have a
111. 3:52list of symbols again just like the
112. 3:55batch importer. So put as many symbols
113. 3:58as you like. And then we have the list
114. 3:59of the time frame. So I have given it 15
115. 4:02minutes, 30 minutes, 1 hour and 2 hours.
116. 4:04And next we have the start date and end
117. 4:06date of the back test. And then we have
118. 4:08the data. So because this is back test,
119. 4:10we don't just have trading routes, we
120. 4:12also have data. So in the example of the
121. 4:14mini version RSI strategy, we had BTCUDT
122. 4:18and time frame of 4 hours as our data
123. 4:20route. So it's also important to put
124. 4:23this here. Now a little note here. What
125. 4:25if we're not trading BTC anymore because
126. 4:26we're giving it a list of symbols,
127. 4:28right? So, what if it is trading ETH
128. 4:29now? Well, when you define something
129. 4:31like this as a data route in Jesse,
130. 4:33something you guys may not know is that
131. 4:35behind the scenes, it actually imports
132. 4:37the data for that time frame for all the
133. 4:40trading symbols. So what I mean is that
134. 4:42if I put this here, even if my trading
135. 4:45symbol is ETH, because I have this BTC
136. 4:47UCT in the 4 hours time frame, Jesse
137. 4:50will also import the data for ETH US on
138. 4:53the 4 hours time frame. So this will
139. 4:55work no matter what. But sometimes the
140. 4:58bigger time frame that you need isn't
141. 4:59for the exact symbol that you're
142. 5:00trading. So you got to be careful with
143. 5:02that. So what if for example, I'm
144. 5:03trading a memecoin, but I also want the
145. 5:05data of solicity in my strategy. That's
146. 5:09a pretty good use case. So in those
147. 5:10cases you need to make sure this exists
148. 5:12here. Now I'm going to comment this out
149. 5:14again. All right. So this is good. Next
150. 5:16we have the config which will get the
151. 5:19starting balance the trading fees which
152. 5:21by the way I have set it a little bit
153. 5:22higher than what it is on the actual
154. 5:24exchange. I do this because I want to
155. 5:27account for the slippage or any kind of
156. 5:30fees that we are not taking into account
157. 5:32because in general I like my back test
158. 5:34to be very strict so that if my strategy
159. 5:37is able to produce profits in my back
160. 5:40test when it reaches live trading and it
161. 5:43pays even less trading fees it has a
162. 5:45better chance. So you hear all the time
163. 5:46that live trading isn't as good as back
164. 5:48testing. So because of that I try to
165. 5:50make my back test as hard as possible.
166. 5:52And next we have the type which is
167. 5:54features. Features leverage is 10. Then
168. 5:56features leverage mode is cross. And
169. 5:58then the exchange. You don't have to
170. 5:59change this because we already defined
171. 6:01it uh above here. And then we have the
172. 6:03warm-up candles which is by default 210
173. 6:06in Jesse. And that's it. We don't really
174. 6:08care about anything else. So let's go
175. 6:10ahead and run this. Now by the way, when
176. 6:11you run this for the first time, there's
177. 6:13good chance that there's going to be
178. 6:15some dependencies that aren't installed
179. 6:16on your machine. But most of them are
180. 6:18exactly what we use in Jesse. And of
181. 6:21course the most important part which is
182. 6:23the back testing itself is using Jesse's
183. 6:25research module. So you need to ensure
184. 6:27you have Jesse installed and some other
185. 6:30libraries that may not be coming with
186. 6:33Jesse. I think seabour is one of them
187. 6:35and maybe also matt plus lib. So if you
188. 6:37ran it and got an error like this just
189. 6:39make sure to run a command such as pip
190. 6:42install and then the name of that
191. 6:44library. Do this one or two times and
192. 6:46you should be good to go. And the other
193. 6:47thing that you need to be very careful
194. 6:49is where to put these files. You need to
195. 6:51put them inside your Jessie project. So
196. 6:53here the bot is the name of my Jessie
197. 6:56project and here's my strategies folder,
198. 6:59the storage and then we have the root of
199. 7:01my Jessie project and this is where I
200. 7:04need to put this. All right. So let's
201. 7:05run this now. First of all, it is saying
202. 7:07base here. So the cond environment that
203. 7:10I have for Jesse isn't activated yet. So
204. 7:12let's do that.
205. 7:15activate
206. 7:17Jesse 3.12 and then I will simply say
207. 7:20Python batch
208. 7:23back test
209. 7:28and notice that it is using parallel
210. 7:31processing to finish these back tests as
211. 7:33fast as possible. Now, while that's
212. 7:35going, I want to quickly remind you guys
213. 7:36about our Telegram. It's the fastest way
214. 7:38to get notified about my future work,
215. 7:40whether it's a new tutorial or a tool
216. 7:42that I create. Also, don't forget to
217. 7:44check out our free Discord where more
218. 7:46than 5,000 members like you and I are
219. 7:48hanging out there and helping out each
220. 7:49other with algo trading so we can all
221. 7:51succeed together. The links for both are
222. 7:53down in the description. There we go.
223. 7:55So, first let's take a look at this. So,
224. 7:58I'm going to zoom out so we can see this
225. 8:01table a little better. All right. So, I
226. 8:03hope this isn't too small for you guys
227. 8:05to see, but basically we have a table
228. 8:07here that shows us the result of the
229. 8:09back test. And by default, it's been
230. 8:11sorted by sharp ratio. So for example,
231. 8:14we can very quickly see that BTC UC on
232. 8:17the hourly time frame produces the best
233. 8:19sharp ratio, which is 1.56.
234. 8:22And next we have ETH USC on the 30
235. 8:24minutes time frame. So that's
236. 8:26interesting. And then we have ETH USC on
237. 8:28hourly. Then we have BTC on 30 minutes
238. 8:31producing the best sharp. Then we have
239. 8:33B&B on the hourly and so on. And we also
240. 8:36have tables. So let's take a look at
241. 8:38this. First is displaying them by the
242. 8:41returns and then by win rate. So we have
243. 8:43BTC on the hourly producing 94% win rate
244. 8:48which is amazing on ETH hourly still has
245. 8:50the best win rate but when it comes to
246. 8:53something such as sharp which is what we
247. 8:55care more BTC USD on the hourly produces
248. 8:58the best results for ETH. However, we
249. 9:00can see very quickly that on the 30
250. 9:03minutes is it has the best results which
251. 9:04is 1.25. And then we have another table
252. 9:07for the max draw down. So let's close
253. 9:08this.
254. 9:10Next, we have another table which shows
255. 9:12us the net profit on different time
256. 9:15frames. So for example, on ETH, I can
257. 9:17very quickly see which one is giving me
258. 9:19more. On BTC, I'm seeing that the
259. 9:21hourly, which is this orange one, is
260. 9:23producing better numbers. On BNB, only
261. 9:26the hourly is profitable and the other
262. 9:28two are actually negative. And on soul,
263. 9:30I can see that almost all of them are
264. 9:32bad. And we have another table for shop
265. 9:34ratio.
266. 9:37And then we have another one which shows
267. 9:39us very quickly the max draw down and
268. 9:41the win rate and also the number of
269. 9:43trades. So the bigger the circle here is
270. 9:46the more trades it is taking. So for
271. 9:48example, this one looks amazing. This
272. 9:51one which is for ETH on 30 minutes is
273. 9:53also pretty good and so is this one. So
274. 9:55I guess this last table is the most
275. 9:57useful one to me. Of course I have the
276. 9:59table of sharp ratios. But you might be
277. 10:01thinking okay so you basically just gave
278. 10:04us the same results. Well, go around
279. 10:06this yourself on as many coins as you
280. 10:08can because I did. And after finding the
281. 10:12best combinations, my results went from
282. 10:15this, which was producing 848%
283. 10:19in 3 and 1/2 years with a maximum of
284. 10:22minus 22. It went up from this into
285. 10:26this. So now on the same time period it
286. 10:30is producing 21,763%
287. 10:35with a max draw down of - 32%. So if I
288. 10:38compare the sharp ratio for example so
289. 10:40you can see it went from two up to 2.7
290. 10:43and even the average holding time went
291. 10:46down from 70 to 43 hours which is
292. 10:50better. So this number is simply crazy.
293. 10:53And to trade it, instead of three
294. 10:55assets, I'm trading eight different
295. 10:57coins with different time frames. And I
296. 10:59found them all using this script. Now,
297. 11:01I'm not going to show you the exact name
298. 11:03of those. It's going to be your
299. 11:04homework. So I gave you the script. Go
300. 11:06run it yourself and find the best
301. 11:08combinations. Now, in case you want the
302. 11:09source code for the strategy I just
303. 11:11showed you, first of all, go watch the
304. 11:12video. It's right before this one on my
305. 11:14YouTube channel. And also, if you want
306. 11:16the source code of it, you can find it
307. 11:17on our strategies page, which is
308. 11:20jesse.trade. trade slash strategies. And
309. 11:22here's the results of that one. You can
310. 11:25check it out the results for different
311. 11:27time frames and symbols to see which
312. 11:30ones perform better. And there's also
313. 11:32some metrics here. Now, this is a new
314. 11:34type of video that I just did. I shared
315. 11:36with you a script that I wrote. It's not
316. 11:38an official feature of Jesse yet, but
317. 11:41it's very much could be. So, if you want
318. 11:43to see this feature, this point screener
319. 11:45inside Jess's dashboard. So you won't
320. 11:48have to run a different script to get
321. 11:49the results, but let me know and I will
322. 11:52build it into Jesse. So if that sounds
323. 11:54good, let me know in the comment
324. 11:55section. Now, if you want to support me
325. 11:56and my work to make more videos like
326. 11:58this, please make sure to give the video
327. 12:00a like and subscribe to the channel.
328. 12:01Thank you so much for watching. I'll see
329. 12:03you in the next one.
330. 12:16[Music]
331. 12:17[Applause]