Take a look here on how the strikes can be generated: https://github.com/aquarians/Public...ians/backtester/jobs/GenerateTestDataJob.java >> I can open source it, perhaps I'll put it up on github later today. "Later today" proved to be "two days later", classic software development estimates, right? Well it's "open source", which means I don't get executed for failing to deliver my "sprint commitments". Main problem though are real sprint commitments which drained me of so much energy that I had no motivation left at end of the day. Thanks God for the weekend! Btw, I'm old enough to have lived in a period and part of the world where "weekend" meant just one day, Sunday. I hated Sunday. Best time of the week was Saturday afternoon. Sunday was too close to Monday to really be of any use
I think you will see a movement from 2 day settlement to 1 day and eventually instantaneous. Please don’t make trading be 24/7. No one wants to be awake at 4am concerned about a trade. Also the off time allows for another type circuit breaker for significant events. I could see expanded hours maybe including after hours as just an expansion of regular trading rather than the separate thing it is now. But there really needs to be an off time. Liquidity will continue to increase some as the lowered and eliminated commission rates are digested and expanded to more people. There will be some adjustments based on the recent GME debacle. Not sure what will happen here but something should be done to prevent a complete blow up of a stock like that. The transition of retail trading will continue. It will likely never replace funds as the primary mover of the market as a whole but it will take its place as a much closer 2nd place. This is more of a wish but I would like voting rights to move from funds to the individuals holding the stock in the fund. Much could change if proxy voting stopped being concentrated in the hands of a small number of fund managers. In reference to your Moore’s law comment, I see an increase of things like wider spread AI software being involved in trading. Maybe to the point of individuals eventually having direct access. I am not sure that I think this is good, just inevitable.
Thanks for posting, I'll look through it more thoroughly later. If you are interested in feedback, I can provide some. It also would have been cool to see the full commit history, to see the progression.
You wrote the entirety of that repo without hitting save until the end?! What I mean is that it looks like there was a lot more changes going on, this is just a copy paste of the source itself, no history. For example, I see a lot of similarities in your code to mine, I'm wondering if we both went down the same route.
Thinking about it you're right. Thinking about it you're facing a killer at least as ruthless as you. WW2, fly over "Pitesti" and bomb them. Fighter pilot, nice to meet you flying over our petrol fields B-)
Answering again coherently after a good sleep. The code I put there is the result of several iterations over the years. Not that it takes years to do it, just slowly adding stuff to a codebase when I got time and mood. What I got there: - Importing stock and option history from files in the format provided by historicaloptiondata.com into a PostgreSQL database - Generating simulated stock and option history data into the database: stock moves by a geometric brownian motion of constant volatility (Black-Scholes model). Can specify a different volatility or a smile for options - wouldn't make sense in reality - "Playback" of historical data imported in the database in either single step or continuous mode. Can be configured to work on several threads. - A very basic calculation of implied volatility and then logging the quality of fit (look for "MODELFIT underlier=.*err=" in the log. Ideally error would be zero if both the data and the fit would be perfect. - Displaying the volatility curve and the option prices along with fair value calculation based on implied vol. Ideally the fair value should lay between bid and ask so there should be no arbitrage opportunities. When fair value lays outside bid / ask then an arbitrage opportunity exists and respective prices are marked with a red background. Example of a perfect fit:
Getting a perfect fit is kinda tricky. One needs to fit not just the implied volatility but also the forward price and sometimes the interest rate and dividend yield. Also rate and yield make sense for stock indexes, if one were to be very accurate, the pricing model for individual stocks would consider absolute dividend payments. Quite a pain to maintain them, but plain Black-Scholes with dumb fitting is good enough if you're not a market maker.