Advance Concepts for Stock Trading !!!
from excel to the broker's API, even that has a bridge? yes you need the bridge, but then some broker will allow you to read through it suppose someone is an engineer or a sophisticated learner I am asking this question for them because this might not be relevant for a common man trading view is done by normal people at least but if someone is sophisticated, so what kind of languages, tools, or software they can use to bridge the gap? that list is unlimited welcome to another episode of Face2Face, I am Vivek Bajaj from this medium of face2face, I am trying to give you the best market knowledge be it trading or investing even in this video, I have come up with the best knowledge especially for those who believe in systematic trading what is systematic trading? it is something when you define that 1+1 is 2 you have defined that you can do this only when this happens somewhere we also map that in our mind but when we are entering the market, the sentiment makes us weak systematic traders or you can say algo traders these people are those traders who tell their computer to not listen to them do not listen to your mind, do only what I have told you to so today we will talk about algo trading, systematic trading and other such concepts and the friend who is here, is a part of the stock market since 20 years he has so much experience in the stock market he has built a whole ecosystem which is commendable, so before I invite him their is a disclaimer my intention in face2face is just one to give you right knowledge I interview a lot of people through this process everyone have their own business model generally, I do not participate in anyone's business model, but when I do I will let you know but from the knowledge point of view, I will give you the best if you are doing any business dealing with anyone then that is your call I don't promote anyone here it was very important to give this disclaimer so let's welcome today's guest and welcome Mr. Rajesh Sriwastava Hello sir, Thank you Vivek ji for calling me so how are you? I am good, very nice! Rajesh ji has come to Kolkata, we will talk about algo trading, systematic trading, and all the concepts that you have followed I will take that out, you know how smart I am in taking things out so first of all thank you for coming to Kolkata, and be comfortable here so 20 years in the market, but algo has just started, so how has your journey been in the market in these years? in the beginning, it was discretionary, then later it was about what are other people saying, for example if someone is talking about a stock 10 people in one room used to trade, used to ask the seniors how were you able to trade this, even at that time they used to do wishful trading, but they were seniors so we used to listen to them, then their was a time when we had to learn this charts, then I started doing technical analysis, and by the time it was 2008, then we removed our accounts then I understood there is something wrong in the whole process and something is just not working, then we focused on data then I thought no more wishful trading, even if I do but it won't be wishful trading, then we went to numbers, and then we did data crunching took a trade according to fundamental, and you used to take both long and short trade, but there are market conditions, no matter how long it is you won't be able to make money you won't be able to make money, so first we quantified the market, and we can't do it wishfully we have to make a trend it should be data-driven or any method of fundamental, up trend we will do long, our work was to change the stock selection method, it should be more than one method so eventually we started building up and that time the data was not that available there were many senior people who were more than happy to help us, so we learnt some number punching from them, then we built a proper model there was a time when we wanted to scale and we needed algo, so we used it twice and then closed it something is missing so you need to walk with people who knows this, so I started learning about it and talked to people even after doing it twice, it was not fun, then I thought I need to bring a professional into the game, I cannot do everything together then I got people from the infrastructure and from tech, so you are doing algo from how many years? this whole algo, if you see on and off then it will be 4 and a half years, so now you don't do discretionary trading at all? no not at all it is all algo driven, the whole strategy? and how many years it has been? 13 in index, and in non-index it is 8, at any point of time are there positions in every strategy if there are conditions meeting then there will be a position, and it is also possible that one strategy is saying long and the other is saying short, most probably it happens most of it is not talking to each other, and it is mutually exclusive, but it doesn't happen like that, there is a time when there is a daily trade in index but in stocks, trades of 2-3 days won't be visible friends, we are going to do a very candid discussion, and you will have fun, you will have to stay till the end we will do an insightful discussion, and we will simplify the concept of algo trading so that even kids will know what is algo trading from this video, so let's talk about the A in excel so we will think like that algo, we have to take things out of our system, so tell me the whole process before we go into the whole discussion so the first thing is to study it because this cannot be discretionary, suppose you have to go along this trend line, the system will not understand this yes there are technical analysis driven algos, it cannot be discretionary it has to be coded, so first cell one you have to quantify so let me give you an example that if I take nifty at 11 am in the day and if I sell it at 12, and I keep doing this, then I will make money right you have reduced the margin, the total margin will be 50,000 rupees, and your drawdown is 25000, and your backtest is doing really great you have done a margin adjustment, taken a trade and your maximum drawdown is tomorrow, now you are left with 25,000, will you able to take a trade the next day so we have to develop a methodology about what all I have to see in this backtest, in that drawdown what is the maximum recovery time period, anything more than 10-11 days is not the greatest it is 40,000 and then the daily average is 2000 rupees, you need another 50 days to gain this money, correct you are doing too much for nothing, okay, and then you are built on that and then you have to optimize if you can sell it at 11 or 11:02 today we will discuss the strategy and try to discuss your parameters we have started an open interest strategy as many people use that first I will say that open interest is fine because there are many strategies built around that, but let's understand open interest neatly we will take out the bug of the thought process because if the thought process has bugs then it will be the same in the code, and it will also have a bug another strategy I have, which I call it, so what is open interest? in FnO there are around 200 stocks which stock we have to trade, we can stop talking about index, retailers like stocks more and not index so open interest is one of the major to track the market activity, which stock has the most activity people will go with that and we are talking about futures, we will talk about options later, now people know about the long build up, price, and open interest goes up so it is long build up but why? we have seen many stocks, some have 2% open interest, 0-0.5% is the gain because people have heard about it from website or some source that OI is up to long build up but if the exchange has opened up then surely there must be some work textbook wise this is a long buildup, you can code this one, and this will go into algorithmic trade but will it give a good result? we will see the side effects, in that 3% it is not much, okay fine we will do it 10%, so this is a decent size, we said it is exchange so some work is happening every work cannot be built up, okay then you have done it 10%, is it good enough, let's extrapolate it in terms of lot, maybe mrf is a low liquid stock in 5 lots also it can be 10%, we can't say the same in the market that this much is being done you have to see where is the traction, the percentage is not the only issue, you can convert in value but you can't say 100 lot is good enough fine we have done it 5, it is still a long build up so here we have done 500 lots so the market has a lot of traction here, technically it should be a long buildup, but it is something to think about, all this was done in 30 mins, but this is not a long build up because they are fighting on their own this is a long build up, possibly this is a fight between bulls and bears, and then everyone is adding their positions so you have done 1700 lots, we have taken yesterday's examples, Siemens 1700 lot is good enough, this is tomorrow's EOD, but we also have to refer to the system to understand the data if we see Siemens and November expiry then 1700 is still there, yes because this is open interest and it is good enough so the net position it is, so it is Siemen's in October it is 1700 lot so nothing is happening, so the model is here in the number of the lot you have to optimize that it should be a COI and not one month, cumulative open interest, in stockedge we show cumulative open interest, the reason is the roll-over problem if I tell you about the size then on any given day it will not size more than that and it is 1700, so it is 3x, which means something big is going to happen, but in net level it is only 15-20 what I mean is that you should keep a threshold, what percentage you will have a clean buildup here I want to tell you something, suppose you are looking at this data at 2, and say the price is 5% on the top, 1700 lot is done you will say it is a long build up it was 10% above, now you are 5% down, there could be 2 reasons, one is you have un-winded or there are new shots so you have to see the journey of the price or the OI, so intraday price movement and OI you have to see together but if it is going from top to bottom then the OI is not changing, most likely these are short positions the theory we use is that the open interest point is where there is a stock, and a seller is definitely sitting if we see the straddle, then the first day of the month is the busiest and it is too premium it will expire as seen on October, and it was 730, and the average price is if you add it then it has comfort zone this is at the combine level, combining all strike price, in call unit of 519 has been removed but it has a single expiry, even if it adds in next expiry it will be a big data seller is not dumb that he will only sell the call because the seller is also selling the put who is selling this? people who want to earn little bit or want in theta decay which is combined premium, so we can say that this one was going to stand in 780 the day when the highest open interest was built, so on the first friday we have kept that set the OI of day after tomorrow we can consider as the base, we will see the expiries of 3 fridays I will take you to the PPT, I am sure there are a lot of things that you have not prepared the pivot chart looks so beautiful, if you see the previous three expiries then there is a marginal call we will back the combined value, and then we will go with it we have used an algorithm here and we have recorded it since last 2 years but if I only focus on these 30 stocks in this expiry, if this is future then you have to focus on 21 stocks then they see the price, and then they see the relative strength then I use some tweaks in relative strength whenever I listen to Relative Strength then my ears become more alert, yes you use it in the application when the OI was building up then the price was going up and down, or it was about the previous season your question was, this has happened, now how to take the trade, Vivek sir, we have done this to generate data points but I can trade any instruments when the signal was generated, to generate signal you need more data points