ACMS > Updated deep thoughts

deep_thoughts.md

@@ -51,4 +51,5 @@ I am trying PGN model now on the stocks. The major problem here is that the rand
 The idea to use use supervised network first actually worked and now PG is also generating profit on the test data. The model is not just holding stocks its improving also.
 
 ### Checking profitablity:
-I have checked different stock brokerage transaction cost in India. The minimum one can go is to pay just STT for brokerage tax which is 0.025% on an intra-day transaction. The rate is 0.017% on all Futures and Options transactions. Now, based on model, I have to write code to check what is the ow much friction costs can the strategy win past. Also, there is one other idea also of separating test data into 4 datas and checking different average profit in 4.
+1. I have checked different stock brokerage transaction cost in India. The minimum one can go is to pay just STT for brokerage tax which is 0.025% on an intra-day transaction. The rate is 0.017% on all Futures and Options transactions. Now, based on model, I have to write code to check how much friction costs can the strategy win past.
+2. Separating test data into 4 smaller sets and checking average profit in them.