Full notes
Full Line Go Up update
Read the full published notes in a cleaner layout. The original post stays linked below.
What changed
- Gameplay
- Balance
- Performance
- Events
Welcome back! Today we will walk through the first of three market conditions included in the demo: Easy difficulty with an Aggressive theme (denoted as Easy/Aggressive).
Okay. So what is the underlying truth in this scenario?
As described in the Dev Diary 1, because the difficulty is easy, the underlying truth depends on only a single stock:
ABC and QE will increase if BCUC increased on the previous turn.
ABC and QE will decrease if BCUC decreased on the previous turn.
The magnitude of the changes for ABC and QE are approximately proportional to the change in the BCUC price from the previous turn.
When playing a game, these relationships may be visible when only the BCUC, ABC, and QE are selected in the figure, although the differences, especially for BCUC, are relatively small.
While the relationship can be identified within the game, it is also visible in the simulation analysis of the strategy creator. For example, the change in BCUC strongly predicts the change in ABC prices.
This market condition has another component: the other stocks vary randomly depending on whether it is a “bull” market (more likely to increase) or a “bear” market (more likely to decrease). Easy/Aggressive tends towards “bull” markets, especially if the prior turn was a bear market. In other words, if a bear market just occurred, then a bull market will likely occur. Otherwise, no other factor informs the likelihood of a bull or bear market.
Note: The concept of a “bull” versus a “bear” market appears in other but not all market conditions. In Easy/Aggressive, it primarily serves to add a rollercoaster feeling. Other market conditions, especially medium and hard difficulties, the “bull” versus “bear” market adds more nuance to the underlying truth by depending on stock prices.
How do we turn the underlying truth into a profitable trading strategy?
A simple but effective trading strategy has two-steps:
Buy ABC when BCUC increased: This captures the upside of the primary signal in the market.
Buy BONDS when BCUC decreased: This avoids any downside by moving to bonds, which have a constant price and 4% dividend on each turn.
This strategy focuses solely on the primary component of the underlying truth. The rules for implementing in the strategy creator are displayed below. Since every component of the strategy is executed sequentially, a “sell all stocks” step is added at the beginning.
This is a good strategy and is the market-strategy AI for Easy/Aggressive. However, the role of “bull” and “bear” markets encourage a second dimension to the trading strategy. Specifically, when BCUC decreases, this is likely due to a “bear” market and the following turn will likely be a “bull” market. Thus, when BCUC decreases, we could buy any stock other than ABC or QE, which will decrease, instead of BONDS. This trading strategy may perform better because stocks are likely to increase more than BONDS, which remain constant and provide a fixed 4% dividend. To assess the likelihood of better performance, we can observe the relationship in the simulation analysis by comparing the change in any stock besides ABC or QE compared to the change in BCUC.
As demonstrated in the simulation analysis, this relationship will likely lead to higher portfolio value. One approach for implementation is slightly modifying the prior strategy:
Buy ABC when BCUC increased
Buy FIN (one of the other stocks) when BCUC decreased
This simple change results in a >30% increase in the average portfolio performance compared to the market-strategy AI.
Is this really an optimal trading strategy?
No, there are smaller optimizations around this general approach:
The initial price could be considered because the same absolute price increase will have a larger proportional impact at a lower initial price. For example, a $10 increase is 10% at $100 and 5% at $200.
Stock dividend could be considered because some stocks have notably higher dividends, which will provide a higher turn-by-turn return if all other things are equal (for example, similar initial stock price and similar expected price changes).
In fact, both of these factors could be considered and could result in a mathematically optimal solution. However, that is a challenge left to the players.
What is next week?
We will cover the medium difficulty market condition included in the demo.
Source
Changelog.gg summarizes and formats this update. How we read updates.
