Rolling averages smooth out short-term fluctuations in data by averaging a value over a defined time window, typically 3 or 6 months, so each new data point is the average of the current and previous months. Analysts often use them to produce smoother trend lines, especially when working with volatile or limited datasets.
We intentionally avoid rolling averages because they can mask meaningful data patterns and create a false sense of trend stability, especially when used with small sample sizes. Most requests for rolling averages arise when users have filtered their analysis down to just a handful of listings, which can lead to misleading conclusions.
Instead of smoothing small data sets, we encourage users to broaden their analysis to include more relevant listings, improving the reliability and economic validity of the results.
While smoothing may make lines look cleaner, it often obscures important short-term market signals, especially in fast-moving or seasonal markets. In cases of low listing volume, the variability seen in the data is not “noise” to be smoothed away; it’s actually a reflection of the real volatility that comes with small datasets.
By prioritizing statistically meaningful groupings, our dashboards aim to help users interpret real trends, not create illusions of stability.
We recommend:
Broadening filters (e.g., increasing geographic area or price range) to ensure a larger and more stable data set.
Not at all! You can still explore niche segments using our filters—but with the understanding that smaller datasets will naturally show more variability. Rather than smoothing that variability away, we want users to interpret it in context, or combine it with wider segments for more robust insights.