[HN Gopher] Zack: A Simple Backtesting Engine in Zig
       ___________________________________________________________________
        
       Zack: A Simple Backtesting Engine in Zig
        
       Author : sksxihve
       Score  : 65 points
       Date   : 2025-04-17 03:36 UTC (2 days ago)
        
 (HTM) web link (github.com)
 (TXT) w3m dump (github.com)
        
       | JSR_FDED wrote:
       | Interesting. This looks like just the thing to try out Zig. Where
       | can I get historical data to try this on? Are there data sets
       | that are pretty recent? I'd like to get a few months of data but
       | right until yesterday.
        
         | jakeogh wrote:
         | Checkout https://github.com/ValueRaider/yfinance-cache
        
       | celltalk wrote:
       | I was actually revoking my hobby site for backtesting and
       | casually developing it these days (indicatorinsights.co). Just
       | saw this, very interesting work!
        
       | henning wrote:
       | It's still not clear to me that past data has any information
       | about future prices or that repeating trends appear with any
       | statistical significance. Especially major indexes that aggregate
       | lots of equities. It's not clear if they ever did and if they did
       | if they still do now that everything is electronic and very fast.
       | I understand there are studies showing momentum phenomena, etc.
       | Maybe there would be if Renaissance Technologies and these other
       | gigachad companies didn't exist, but they do.
       | 
       | I understand that "stocks in play" (boomer slang for meme stocks)
       | display whacky irrational behavior that makes no sense, or there
       | is just a temporary combination of high volatility at relatively
       | low market cap, and therein lies potential profit opportunities.
       | It's not clear whether an algorithmic trading strategy is
       | applicable to such situations although I do see people always
       | mention "stock screeners" which seems to just be code to filter
       | time series.
       | 
       | After taking into account transaction costs, slippage and taxes,
       | it's not clear that there is anything better to do than buy and
       | hold low-cost index funds.
       | 
       | Also, there is a separate problem around having the computer
       | structure/segment large trades to reduce market impact to get the
       | best deal possible by looking at the order book to wait for
       | people willing to buy/sell a lot at reasonable prices or
       | something -- that is a different kind of algorithmic trading from
       | what most people mean when they say they are backtesting a
       | trading strategy like moving average crossover or something.
       | 
       | It all just feels too much like alchemy. But, if you become
       | independently wealthy by writing code or you just don't lose a
       | lot of money and you find it fun so it's sort of like paying to
       | do something you enjoy, God bless.
        
         | code_biologist wrote:
         | _it 's not clear that there is anything better to do than buy
         | and hold low-cost index funds._
         | 
         | Sure, 100%. The growth of this strategy has limitations though.
         | Check out Mike Green's interviews on Youtube about the growth
         | of passive investments post 2008 as a source of systemic
         | volatility.
         | 
         | Though I can't tell if this project supports it, how do you
         | feel about non-stock price indicators? Things like housing
         | starts, gas prices, or various constructions of a "Walmart
         | indicator"?
        
       ___________________________________________________________________
       (page generated 2025-04-19 23:01 UTC)