Python 股市Screener 分享

簡單來說就係點樣有系統可以跟而做到「贏要谷輸要縮」

如果係我個screener第一個條件都唔得 要$10以上

Survivorship bias free既data多數都要用錢買

其實你要screen呢啲有現成既
用thinkorswim就得

用這個talib 會行得快好多。記得os 要另裝talib和裝python 嘅pandas_ta pkg。本身default 行MP。

https://github.com/twopirllc/pandas-ta#indicators-by-category

多謝樓主分享
有個小小嘅建議，可以試下用yf.download呢個function 黎一次down 晒啲data落黎先慢慢filter, 應該會快啲。另外，可以將啲filter過程變做一個function先再try exception，咁可以唔使每個variable都要做一次。

ching
這個是否vcp?

可以呀
但止蝕位有少少遠 13.31 /12.26 如果佢pull back到 13.80-14.00 就正

ching
想請教點睇止蝕位？

因為只買短期突破 所以都只留意最近幾天既最低位作參考
另外止蝕位最多容許5-6%就算 咁你突破後10-12% 都有2R 你有4成以上勝率都ok

謝ching

感謝你

ching,
請問有冇人用過facebook 既 Prophet 做forcast? 有無用?有幾準?
我想試下用黎測量大市變化，例如某個板塊，上升趨勢嘅股，資金流向等。有冇其他建議。

https://facebook.github.io/prophet/
https://towardsdatascience.com/forecasting-stock-prices-using-prophet-652b31fb564e?gi=acfc10545a88

喂 師傅 比個requirements.txt好不好

我

https://www.ijcai.org/proceedings/2020/0641.pdf

read this paper.

This manages the portfolio instead of playing with one stock only.

The bad thing is during the whole period, the portfolio selected is fixed.

forgot 不過github有record好似 得閒揾揾 記得係唔差

(vcp) ➜ vcp_screener-main python vcp_screener.py
File "vcp_screener.py", line 138
if os.path.isfile(f'{daily_file_name}.pdf'): #萬一出現重複值, 自動continue 跳出迥圈, 回到上一個try,向前找出未下載資料既weekday
^
SyntaxError: invalid syntax



樓主用緊python幾 ？
我裝咗3.8 run 唔到

推一推
新增Industry group ranking, EPS rating

def get_tickers():

filter = ['ta_sma200_pa', 'ta_sma50_sa200', 'ta_sma50_pa', 'sh_price_o10',
'sh_curvol_o2000']
stock_list = Screener(filters=filter, table='Performance', order='Ticker')
df = pd.DataFrame(stock_list.data)
df.set_index('Ticker', inplace=True)
df.drop('No.', axis=1, inplace=True)
tickers = df.index.to_list()

with open('stock_tickers.pickle', 'wb') as file:
pickle.dump(tickers, file)

return tickers

樓會唔會試下用finviz.screener做一次initial screen先 咁樣已經screen淨幾百個tickers
會唔會快啲?

利申 新手想學野

想請問free 裡面可唔可以extract 埋啲non-market 嘅data?
姐係revenue ebit ebitda 呢啲？
好似CIQ 咁