Scoring Post-Event Price Recoveries with Limited Historical Data
Summary
The document frames a selection problem in an online game’s item market: prices fall when a double-experience event increases supply, and the trader wants to buy during the event and sell about a month later. The proposed starting point is to compare each item’s post-event price decline with its price around a month afterward. The question asks how to turn this pattern into a score that can be applied semi-automatically across items.
The main analytical issue raised is that items may reach their post-event lows at different speeds, so measuring the same fixed interval from the event can misclassify their recovery potential. The available history contains only one such event, which gives little basis for distinguishing a repeatable seasonal effect from an item-specific or one-off move. No scoring formula, tested results, or completed strategy is supplied; the document is a request for criteria. Its setting is a game market rather than a financial exchange, and it assumes items remain liquid with adequate supply and demand.
Key ideas
- The proposed trade buys items during a supply-driven seasonal event and sells about a month later.
- A score could compare each item’s event-period price decline with its later price recovery.
- Items may reach their lowest prices at different times, making a single fixed measurement window unreliable.
- One observed event provides limited evidence about whether the pattern will recur.
Tags
Full text
# Decision criteria after seasonal drop (GAME) # Decision criteria after seasonal drop (GAME) I'm playing a game in which you can buy and sell items (it's an mmorpg). Now, after certain events, there is a huge drop in the price of certain items (there is a seasonal double experience weekend in which many people grind a lot, causing the market to overflow with certain items). Now, I want to decide which items I could buy right after that double experience weekend to sell a month after. I have the average price of the item per day readily available (I downloaded all the data to my computer) and it can look for example like this: As you can see, there is a huge drop around the beginning of october after which it will steadily raise again. I only have the data of one of those double experience weekends though... Now, my question is: how can I decide whether an item is a good one to buy in a (semi-)automatic way. Are there any good criteria? I want to calculate a 'score' of an item to decide whether it's a good one or not (let's say that every item is always traded and there is always a sufficient amount of offer and demand). At the moment, I'm looking if a price of an items drops during the days after the double experience weekend and compare it to the price a month later. This still has some errors though, because some items do not respond as fast as other (in their drop). Any theories I should be looking at? Any criteria to investigate? Keep in mine I want to buy DURING the double experience weekend and sell about a month after. Thanks for your time already! PS: I also don't know if this is the correct section for my question...
Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.