65% of Hacker News Posts Have Negative Sentiment, and They Outperform
Updated 15 March 2026
Contents
Key takeaways
- In our Hacker News sample, negatively framed posts averaged 35.6 points against 28 overall, linking criticism with higher engagement.
- Nearly 65% of sampled posts registered as negative, and six sentiment models retained a negative skew despite differences in classification.
- The analysis cannot establish whether negative framing causes attention or whether controversial topics attract both criticism and engagement.
Negativity Bias and Engagement on Hacker News
This Hacker News sentiment analysis began with a simple observation: posts with negative sentiment average 35.6 points on Hacker News. The overall average is 28 points. That’s a 27% performance premium for negativity.
This finding comes from an empirical study I’ve been running on HN attention dynamics, covering decay curves, preferential attachment, survival probability, and early-engagement prediction. The preprint is available on SSRN. I already had a gut feeling. Across 32,000 posts and 340,000 comments, nearly 65% register as negative. This might be a feature of my classifier being miscalibrated toward negativity; yet the pattern holds across six different models.
Six-Model Sentiment Comparison: Transformers vs LLMs
I tested three transformer-based classifiers (DistilBERT, BERT Multi, RoBERTa) and three LLMs (Llama 3.1 8B, Mistral 3.1 24B, Gemma 3 12B). The distributions vary, but the negative skew persists across all of them (inverted scale for 2-6). The results I use in my dashboard are from DistilBERT because it runs efficiently in my Cloudflare-based pipeline.
What counts as “negative” here? Criticism of technology, skepticism toward announcements, complaints about industry practices, frustration with APIs. The usual. Technical critique reads differently than personal attacks; most HN negativity is substantive rather than toxic. But, does negativity cause engagement, or does controversial content attract both negative framing and attention? Probably some of both.
For a different application of sentiment analysis, I tested news sentiment as a trading signal.
HackerBook Dataset: Cross-Validation With 22GB of Hacker News Data
Related to this, I also saw this Show HN: 22GB of Hacker News in SQLite, served via WASM shards. Downloaded the HackerBook export and ran a subset of my paper’s analytics on it.
HackerBook is a single static snapshot (no time-series data). I could not analyze lifecycles, predict early velocity, or fit decay curves. I could compute distributional statistics, inequality metrics, and circadian patterns.