Show HN: When is good time to submit a story on HN?

  a counter that increments with each story upvote

Probably should increment based on each story submitted.

You mean instead of "seconds since this story was posted" it should use "number of stories posted since" ?

Sounds like a pretty good plan.

Though as other people mention you could also add pageviews in the mix, as well as "number of comments since". I'd say a nice linear combination of all three, softmaxed between some limits, added to the number of seconds. The latter bit because you wouldn't want things to go crazy when a huge amount of pageviews or things are posted.

But it's a good idea. Especially since I'm in a non-American timezone, which means that if I want a story to be seen I need to take into account US office hours as well, and your suggestion might sort of take the timezone effect out of it.

This could cause the unintended effect of making it impossible to get to the front page when there is heavy activity. It would cycle quickly between new articles. There needs to be a fixed time element.

Brilliant work, dont suppose you'd reveal (if you have one), your @twitter handle?

It's a nice visualization.

One more suggestion: Do label the time zone. even if it correctly detects the user's time zone, that's unexpected on the internet. Plus you don't label 'current' time... for all I know it's a Time Zone ahead but only updates once an hour. Anyway, all confusion goes away with a label.

Edit: Took out references to my personal time zone

You should put the blurb detailing the formulas under the graph.

On my Nexus S I couldn't scroll at all in either direction, and the right side of the graph was cut off.

Slightly relevant data point of approximated anecdotal evidence: I submitted a "Show HN" for my Android app at about 2 PM on a weekday and got 1 or 2 upvotes within about an hour, so I deleted it and resubmitted it later at about 9 PM (exact same title and everything - I felt bad about resubmitting it but I was really disappointed, sorry if this was a faux pas) and it hit the front page within 5 minutes and hit #1 in about half an hour and then stayed there for 12+ hours. Pretty astounding difference...

Can you move the legend outside of the plot area?

I see that you use Flot, it's quite easy. Just add:

     <div id="legend"></div>

and then when you pass settings to Flot' plot function, add this to the settings dict:

     legend: {
        container: $("div#legend"),
        noColumns: 3, // add this if you want horizontal legend
     },

I agree with the emphasis (link) on "bad"/"good" being confising, and agree with moving the link to "submit", and actually having an explination behind bad/good.

Still nice despite the catch22 though.

I think it's more related to North American lunch breaks and end-of-the-day? I'm on HN now because I take Fridays off for research, but otherwise not much time during the week except for breaks and after hours. I'm not sure how that will change knowing something we all know..

>You should label and explain the axes and sliders. I have no idea what the numbers mean, or what the "pickup ratio" is.

Left-click on the graph.

That would be a good thing for HN, albeit to the detriment of the app in question and its early adopters. But even if there was less ability to "game"/"hack" the system, a public service like this is necessary to ensure that situation remains.

Not a catch-22, it's a game, in the game-theoretic sense.

You're choosing a best-response to everyone else's selected post time. Publishing this post data should make the system approach equilibrium more quickly, where in equilibrium the best post time should be fairly evenly distributed.

I think you have to consider that not all submissions are equal. Some stories are likely to shoot to the top of the front page regardless of their competition. For those, you would want the post to peak when there are the most readers, which requires posting during competitive slots. But for more run-of-the-mill submissions, posting during a time of less competition is probably a bonus.

I know, I know. It's amazing how many times I broke this rule. I do better in future.

Somehow I think there should be a different design rules for labeling "persistent" (?) graphs, i.e. graphs that you look at regularly. If you look at a graph couple times a day you need to know the meaning of axes only one time. After that, the labels are waste of the real estate. I will look at the vital signs equipment to see how it's done.

I still don't know what does the vertical axis represent. Normalized points? At what time?

yeah - that was way too small for me. I had to zoom a lot. For a small project like this it's sort of a "who cares" decision, but I'd hate for that size to be a design "norm".

are you aware of http://hckrnews.com/ ?

Could the author identify him/herself? I think this is brilliant, and would like to web-stalk/follow whom ever created this. In a purely selfish-selfimprovement manner. :)