API pricing model feasibility

So I arbitrarily assume that I have a strategy that gives me 50% return per year.If I invest 1 lkh I get 50k as profit at the end of the year.
Assuming I have taken average 40 trades a month(2 trades a day minimal).In total 40*12=480 trades

Now this 50k is inclusive of tax + brokerage services + API cost of Kite.

The cost structure after one year

Strategy Interest earned =50000
Brokerage cost=480*40(both for buy and sell side)= 19200
Api Cost=2000*12 = 24000

Total Profit end of the year = 50000-(19200+24000)=6800(including tax)

So even if my strategy has a good sharpe ratio and 50% return I lose out on the operational costs.

Now I cannot ask for reduction in brokerages.It is way too small.Therefore my question comes to the API costing. Is it feasible? Also why is the API costs not usage centric?

P.S- This is real life scenario of the strategy.Even after 50% return I am not able to productize the strategy.
  • nithin
    nithin edited December 2016
    Hey Gully, Donno if this is a right way to look at it. If say I am investing 50k to buy my computer to trade, and I have 1lk to trade with, that is like 50% that I have to recover only for computer. If Internet is 2k per month, 25% annually of my capital only for this.. and etc...

    Trading is just like any other business, makes sense only at a certain scale. Do you know that in US, people are not allowed derivatives trading with less than 25,000 USD in the account (Rs 17lks). If you are looking at absolute returns with capital of Rs 1lk in your account, I don't know if you should spend 2k a month for the API if the trading capital will remain the same forever. Charging based on usage will probably end up costing more, and we currently don't have the bandwidth to build the utility to charge based on usage.

    Btw consider this, before Zerodha the brokerage would have been multi-folds of what you are currently paying. Just the cost of live data feed is upwards of Rs 3000 per month for all exchanges. Since you are getting this data within a trading platform, and you are our trading client, you don't have to pay anything extra.

    So yeah, you have to look at all your running costs with a different perspective.
  • gully
    gully edited December 2016
    @nithin thanks for the perspective. I just wanted to appreciate that what you guys are doing with Zerodha is app-laudable and I didn't ask the question with any malice or with skepticism.
    From trading as a business perspective I am just looking ways for my break even cost with algo trading. And therefore my concern and proposal if there is an alternate way to solve this.
    My 2 cents here is that a strategy with 50% return would require a lot of capital(with applied high risks) to make a decent profit out of it
  • arkochhar
    @nithin I am facing similar issue here. Could I make a suggestion that if Zerodha could split the package into two. That is, one for trading and one for historical data feed. Currently, it is that you subscribe to basic trading package which is 2K and then add on the historical data package which is another 2K. People like me who wish to access only historical data for the purposes of backtesting, will have to shell out 2K for services that they will not use, at least immediately. Let me know what you think?
  • amitgaiki
    @nithin is it possible to open up the historic data api for a few stocks for free so that developers can do some testing before subscribing? Also, it would be great if the historic data api is available without having to subscribe the trade api. Consider these as feature requests and let me know your thoughts on it.
  • sujith
    We are thinking about separating trade API and historical data API.
  • joy
    @sujith any chance of a only trade api, like :
    • Trade API - Free, user only pays brokerage :wink:
    • RT API - Costs
    • Candles API - Costs
  • sumitsomani
    @gully A suggestion as it seems your strategies are not making a lot of profit and that is the problem. 480 trades and 50000 rs = 104 rs strategy... out of which ~38% is going to be brokerage (even higher if I consider taxes).

    So boom your 50% return went down to ~31%

    Now 24000/yr cost = 48% of your annual profits. ... Boom you are down to just 6.8% return.
    But if you scale 5 times you could still end up with 26% decent return
    If you scale 10 times you could still end up with 28-29% return

    However in case you are an individual trade my suggestion you should look at increasing profit per trade -> Choose fewer trades because even if you scale brokerage would kill you

    @joy great suggestion bro..
    One for historical prices too -> I think the benefit of having historical prices separate is for first 2-3 month of strategy development traders lose out on 2k extra on trade APIs which cant start using as of now. Once a good strategy is ready, its game on. Initial 6-7k might be a blocker and may hurt adoption for Zerodha.

    @sujith you should def look into it
  • fa7c0n
    @sujith I am also having a similar problem, All I need is the access to historical data to start building and back-testing my algorithm. But having an addition 2K added to a service which I am not going to use in the intial 3-4 months period is just waste of capital. So is there any update on splitting the services? if yes can you point me to relevant documentation?
  • sujith
    Hi @fa7c0n,
    Since this is a developer forum, pricing related discussions are a little out of the scope. Those are done by the business team. Please post your queries regarding pricing to talk(at)rainmatter.com
This discussion has been closed.