What changes will I make to my Manifold search function and trading dashboard by the end of November?
6
203
3.7K
resolved Dec 1
Resolved
YES
Allow markets to be filtered by "closed and not yet resolved".
Resolved
YES
Add a sort/filter for total traders
Resolved
YES
Add a "not" option in the search fields, to exclude a certain term.
Resolved
YES
Allow searching with regular expressions.
Resolved
YES
Update the URL live as the user modifies their search, so that they can copy-paste the URL if they want to send a link to link the search to someone else.
Resolved
YES
Make the "time until close" display update automatically, so there's a live countdown.
Resolved
YES
Fix the lag, so that user input is responsive.
Resolved
YES
Use a market's lastUpdatedTime to determine how frequently it needs to be queried for updates.
Resolved
YES
Listen to /bets and get a new probability for any market that was just bet in.
Resolved
YES
Crawling the full dump of all markets can happen concurrently with other market updates rather than putting them all on hold.
Resolved
YES
Filter by market type.
Resolved
YES
Add a column for volume
Resolved
YES
When the user hovers over a topic, display the total number of markets in that groups
Resolved
YES
Stream data retrieval from the OPFS, so it loads in chunks rather than blocking the page for ~500ms.
Resolved
YES
Let people link to specific visibility options and sorting order via URL parameters.
Resolved
80%
Retrieving new markets should happen concurrently rather than one at a time.
Resolved
NO
Add a sort/filter for elasticity, limit orders, and/or some other similar measure
Resolved
NO
Add a filter to only show markets the user hasn't seen before.
Resolved
NO
Add a filter/sort by recently subsidized
Resolved
NO
Hotkeys for quickly placing bets

I would like suggestions on how to improve my third-party Manifold search page and trading dashboard. If there's something you'd like to see added, add it as an answer, and explain in the comments why you think it would be useful. If you have a suggestion that includes multiple independently-implementable parts, err on the side of making it multiple separate answers.

I am more likely to implement changes that are:

  • Useful for practical purposes.

  • Wanted by more than one person.

  • Easy to implement.

That's not to say I won't take on harder or more niche projects, espetially if it's something I myself would find useful. I enjoy programming, I just don't have infinite time. Please suggest even minor improvements that you'd like to see; the worst case scenario is just that I decide not to do that.

If you see a suggestion from somebody else that you'd also find useful yourself, please like that comment or indicate your agreement in some other way, so that I can get an idea of what to prioritize.

Any time I impliment one of the answers, I'll resolve it to YES. Any that I haven't fully implimented at the end of November will resolve to NO. If there are redundant or overlapping answers, that's fine, they'll each resolve appropriately. If anyone adds an answer that's already been implemented, I'll resolve it N/A.

I won't bet NO (so you don't need to worry about me choosing not to impliment something in order to turn a profit), but may bet or sell YES.

Get Ṁ1,000 play money

🏅 Top traders

#NameTotal profit
1Ṁ71
2Ṁ53
3Ṁ36
4Ṁ9
Sort by:
bought Ṁ2 of Include unlisted mar... NO

I made a lot of bets; tried looking for a few of the features on the current site and couldn't find most after a bit of search (lol), and it's december 1 in the US so I bet a bunch of NO on most things; not sure if there's a private version that's different. the github also seemed not to have been updated in a way that indicated lots of these resolving yes. I'm also betting NO on some of these because they seem like way too much work to implement for too little benefit; for example "Add support for other platforms. Polymarket, Kalshi, etc." would be great, but seems like it would take a bit to figure out, and would be used only very rarely, etc.

Has anyone ever used the mixed sort option? I'm considering removing it.

Blank range endpoints now default to the furthest possible, so you can enter -4,96- to see markets that are within 4 percentage points of either extreme.

I also made relative timestamps allow the negative symbol, so you can input -1d:0 to see any market that was created in the past day.

The numeric fields have been really fiddly to get working both consistently and intuitively, so let me know if you come across anything else that seems like it should be doing something other than what it's actually doing.

We're on Github! Feel free to submit new features and/or point out just how terribly the current ones are implemented.

https://github.com/KingSupernova31/manifold-search

Retrieving new markets should happen concurrently rather than one at a time.

Resolving to 80% since it's depending on how many there are. If there are just 2 new markets they'll be gotten one at a time, but if there are 100 it opens multiple connections in order to get them downloaded faster.

Here's a different market structure where incentives are more aligned with "bet YES if you think the feature would be good". Please submit suggestions!

bought Ṁ1,000 of Add a column for volume YES

People get notifications whenever answer resolves, right? So I don't actually need to post a comment in addition?

i'm not of a conspiratorial mind, but ;P

I swear I'm usually well calibrated!

@RobertCousineau Sorry! I've been focusing on ones that encounter myself, and outdated market data kept popping up in my searches.

bought Ṁ1,000 of Add a "not" option i... YES

Text input fields now support negation; put a minus sign in front of the search term. Works with both regular searches and quoted strings. For example,

-sam altman

Will return only markets that include "altman" and do not include "sam".

-"sam altman"

Will return only markets that do not include "sam altman".

-"sam -altman"

Will return only markets that do not include "sam -altman".

bought Ṁ1,000 of Allow searching with... YES

I've added a custom search field where you can input any code you want. For example, here is a search for all markets whose title contains their creator's username. See the documentation for the details.

bought Ṁ100 of Allow markets to be ... YES

Changes:

  • Dramatically improved the speed of the search and sorting; search results should now be instantaneous as you type and edit the display options. (I'm aware that there is still some lag while the page is loading; that's next on my list of things to fix.)

  • Added the ability to search by whether a market is resolved or unresolved. (For example, here is a list of markets that are closed but not resolved, with the ones that have been closed for longest at the top.)

  • You can now view and navigate though a market's groups. Click on a group in order to see all other markets in that group.

  • Added the ability to sort by creator, if you want to see all markets from a single creator next to each other. And just like with groups, you can click on the creator username to see all their other markets that match your other search criteria.

  • I added an option to display a market's probability visually with a colored red/green bar. It doesn't look great aesthetically, but it's a much more efficient way to quickly see where a market is than having to read a number.

  • The close date column now updates live and will give you a second-by-second countdown.

  • I added an export button to to allow an easy way to grab a list of market URLs or IDs for further use.

@IsaacKing Looks like the unresolved / sort by close time is sometimes returning resolved markets. The third market here (Norwegian LVT) resolved N/A a long time ago:

@EvanDaniel Fixed, should be accurate now.

Right now the tab consumes a huge amount of resources. I've seen over 12GB memory usage, and this is normal:

Firefox 118.0.1 on Ubuntu, seems like on some simple searches it stops doing that and returns to normal, but mixed sort weights keep doing that for an extended period.

@EvanDaniel The CPU usage issue is partially fixed; it will still use a lot of CPU while the page is loading, but will quiet down after that. I'll try to fix the loading soon.

The large memory usage is by design, it stores market data locally to make the search lightning fast. Does that cause a problem for you?

Actually I seem to have made it worse somehow, whoops. Will fix tomorrow..

@IsaacKing There we go, fixed.

@IsaacKing Thank you, that seems much better!

Using some memory is fine, but it was sometimes getting to be well over 50% of my total RAM just for that one page, and there are only ~ 37k markets. Hundreds of kB per market seemed more likely to be a bug than doing something useful.

@EvanDaniel In theory it should be in the few hundred MB range. If you see it get over 1GB again, let me know, that's probably a bug.

@IsaacKing Thanks, I'll keep an eye out! A couple hundred MB seems much more sensible and wouldn't cause any hiccups on my system.

I'm guessing this is the right spot for bug reports / requests?

When entering a URL with mixed sort weights, the "sort by" dropdown does not select "mixed", which means the weights are not shown. So to adjust manually I have to edit the URL again or re-enter all of them after selecting "mixed".

Is there a way to filter markets >x% or <y%?

@firstuserhere Yep, the percentage field can do that.