Haha, in the heyday of Runescape it used to be possible to simply have buy and sell orders sitting in the GE for the same rare items simultaneously, and over time they’d both get filled for a profit. I guess you would call it “market making”.
This is actually kind of how the US market works with essentially 1 market maker who is by far the largest and who also happens to be the largest political donor (100 million a year to gop alone). Oh, and like Bernie Madoff, the owner of that market maker also owns a hedge fund (which like Bernie's makes completely impossible profits year after year). The hedge fund is of course a totally seperately run business from the market maker /s /s /s
So that actually sounds quite realistic to me from the market maker point of view of how things actually work in the US.
This reminds me of when World of Warcraft had a mobile app that allowed you to both browse the auction house and list or buy items. The API was pretty trivial to reverse engineer, making it easy to write an auction house bot. I never got as sophisticated as any ML algorithm, just snagged grossly mispriced items and bought/sold based on the last week's price range; it was mostly about volume. Made a tidy profit, bought my 1mil gold mount then gave up. The app was taken down soon after, no doubt in part because I'm sure I was not the only person doing this.
Sorry if I’m just bad at reading comprehension, but what exactly is being modeled here? It sounds like the author is just getting price data not unlike a candlestick chart. What I don’t understand is if the trading is happening within the 5 minute windows he mentions or if model is finding patterns across timeframes?
So I don’t know this stuff so I’ll stochastically parrot a reply here of my understanding of what I just read.
The ML model is trained on the price spread and volume data collected with the cronjob every 5 minutes and stored in a database, and the trades with their respective data, but on previously collected data from 63 days up to 14 days earlier and then validated on the remaining later 14 days of data, apparently to not have temporal leakage which I googled and means unintentional backflow of training data which screws up the prediction. So to answer your question the model finds patterns over multiple weeks of data, to rank the best trading choices in the present moment. Then the java component executes the trade in-game with the player character.
What I personally don’t understand is if this data collection and training is rolling continually, or if he let it be trained one shot and then it’s smart enough to remain profitable. Don’t entirely understand the 49 day training, 14 day validation thing either.
Haha, in the heyday of Runescape it used to be possible to simply have buy and sell orders sitting in the GE for the same rare items simultaneously, and over time they’d both get filled for a profit. I guess you would call it “market making”.
This is actually kind of how the US market works with essentially 1 market maker who is by far the largest and who also happens to be the largest political donor (100 million a year to gop alone). Oh, and like Bernie Madoff, the owner of that market maker also owns a hedge fund (which like Bernie's makes completely impossible profits year after year). The hedge fund is of course a totally seperately run business from the market maker /s /s /s
So that actually sounds quite realistic to me from the market maker point of view of how things actually work in the US.
This is actually how markets are made - do you think a second hand book shop operates any differently?
This reminds me of when World of Warcraft had a mobile app that allowed you to both browse the auction house and list or buy items. The API was pretty trivial to reverse engineer, making it easy to write an auction house bot. I never got as sophisticated as any ML algorithm, just snagged grossly mispriced items and bought/sold based on the last week's price range; it was mostly about volume. Made a tidy profit, bought my 1mil gold mount then gave up. The app was taken down soon after, no doubt in part because I'm sure I was not the only person doing this.
Sorry if I’m just bad at reading comprehension, but what exactly is being modeled here? It sounds like the author is just getting price data not unlike a candlestick chart. What I don’t understand is if the trading is happening within the 5 minute windows he mentions or if model is finding patterns across timeframes?
So I don’t know this stuff so I’ll stochastically parrot a reply here of my understanding of what I just read.
The ML model is trained on the price spread and volume data collected with the cronjob every 5 minutes and stored in a database, and the trades with their respective data, but on previously collected data from 63 days up to 14 days earlier and then validated on the remaining later 14 days of data, apparently to not have temporal leakage which I googled and means unintentional backflow of training data which screws up the prediction. So to answer your question the model finds patterns over multiple weeks of data, to rank the best trading choices in the present moment. Then the java component executes the trade in-game with the player character.
What I personally don’t understand is if this data collection and training is rolling continually, or if he let it be trained one shot and then it’s smart enough to remain profitable. Don’t entirely understand the 49 day training, 14 day validation thing either.