Hacker News Headlines (game) | Hacker News Reader

Hacker News Headlines (game)

projects.peercy.net

I think you need to adjust the variation of scores, I got 1650 actually guessing, then realized most of the scores were low and got 1800 by just always guessing 239

Ah! Good idea. Hurts replayability though, since we probably get not more than one 1000 pointer each week.

Just change the range of scores. If you're not going to deliberately weight the set of stories to include outliers, then the whole game is really played in the 150-400 range anyways, so make that the slider.

Include some 0 to 4 score posts with interesting headlines. Should be pretty funny I think :)

For example:

Rubin Observatory Discovers Surprise 'Tail' on Iconic Galaxy - 2

Vera Rubin's primary mirror gets its first reflective coating - 157

I got 2038 by just going with the default 373 for all 5

First try:

Hacker News Headlines

Score: 2,003 / 2,500 pts

R1: 684 (guessed 429) - 0.63x - 333 pts https://news.ycombinator.com/item?id=45722639

R2: 298 (guessed 239) - 0.80x - 413 pts https://news.ycombinator.com/item?id=45783074

R3: 273 (guessed 239) - 0.88x - 445 pts https://news.ycombinator.com/item?id=45736461

R4: 288 (guessed 479) - 1.66x - 321 pts https://news.ycombinator.com/item?id=45589715

R5: 244 (guessed 249) - 1.02x - 491 pts https://news.ycombinator.com/item?id=45602124

Play at: https://nbanmp.github.io/claude-web/projects/hn-oracle/

Anchoring!!! If the range has nothing to do with the items in question, just have simple number input. Tested multiple times and all items were below default / midpoint.

ui idea: show the distribution over the slider

I average a score in the range 1500-2000 (out of 2500).

I imagine those who spend more time on hn will ace it.

I got 2000ish, but I remember seeing 3/5 headlines. Perhaps I should spend less time there...

It helps a lot just to have an intuition for what a big article is here. The slider starts at ~1500 votes, which is an insanely successful story. You can get a strong score just by guessing 400 for everything, nudging up +100 for things you remember being popular and down -100 for things that seem obscure.

Yea - I didn't think too long on how to calculate points well. Is there a more balanced strategy?

Identifying some of the truly high-scoring articles (>1000) should be rewarding, and 200 vs 400 points is a pretty big difference on hn

A "replacement-level" front page story is ~200 votes, +/- 50. If you're drawing just from the front page, and not from a deliberately weighted set of successful vs. marginal posts, most stories should be below 400.

(It's weird to say this but I'm not nerdy enough to have actually worked this out with data; it's just intuition from spending time here.)

Could split the stories into buckets and then randomly sample from each bucket. Most stories are small, so they’re currently overrepresented in the sampling.

Why is the score range 239-1000?

Reply on news.ycombinator.com