Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Tweak recommendation difficulty #11

Open
Tillerino opened this issue Oct 15, 2014 · 7 comments
Open

Tweak recommendation difficulty #11

Tillerino opened this issue Oct 15, 2014 · 7 comments

Comments

@Tillerino
Copy link
Owner

Introduce an option which allows to tweak difficulty of recommendations.

@otoed1
Copy link

otoed1 commented Oct 15, 2014

List of possible tweaks: AR value,
OD value,
CS value,
(not sure if this is possible)Circle Density,
Star Difficulty,
PP worth,
Failure rate( I know that airman is statistically the hardest map, due to the most failed plays, this could use percentage values to recommend these maps, this however does not seem particularly useful.
Or, a static difficulty tweak for all recommendations.

@Sellyme
Copy link

Sellyme commented Oct 16, 2014

Would definitely like to see this - Tillerinobot gives me ~125pp maps when I have a dozen >130pp plays and a 150pp, which means that I'll hit pp saturation fairly soon and only progress if I fluke ridiculously good accuracy - having a way to get recommendations about 10pp harder than the usual ones (in terms of speed and aim, not accuracy, as most people aren't near the upper limit for acc on their top plays) would be a very useful way to both find fun maps that you can practice on, as well as getting maps you can try to fluke a monstrous play out of.

I don't believe that making the tweaks dependent on AR, OD, or CS is a good idea, as they're all traits that have little to do with a map's actual difficulty. A 220BPM AR8 OD7 CS4 map with massive jumps and streams is going to be much harder than a 160BPM AR9 OD8 CS5 map that's entirely single-tapping sliders. Star difficulty/pp value is pretty much the only way to really get a "harder" map reliably.

@otoed1
Copy link

otoed1 commented Oct 16, 2014

Sellyme, I was thinking of having an option to prefer recommended stats, such as AR or OD. For example, i'm not very good with low AR. If I want to practice it, I could give Tillerino a command that tells it to only recommend me maps that have an ar < 9 or perhaps even ar = 8. This would be separate from the actual difficulty scaling, as you said selecting maps that are 10 pp or so higher than what the game is recommending you. Maybe that idea should go in a different issue...

@Tillerino
Copy link
Owner Author

I don't think I'll ever add filters to Tillerinobot and I've been saying
this for a long time (see FAQ).

On Thu, Oct 16, 2014 at 4:49 PM, otoed1 [email protected] wrote:

Sellyme, I was thinking of having an option to prefer recommended stats,
such as AR or OD. For example, i'm not very good with low AR. If I want to
practice it, I could give Tillerino a command that tells it to only
recommend me maps that have an ar < 9 or perhaps even ar = 8. This would be
sperate from the actual difficulty scaling, as you said selecting maps that
are 10 pp or so higher than what the game is recommending you.


Reply to this email directly or view it on GitHub
#11 (comment)
.

@otoed1
Copy link

otoed1 commented Oct 16, 2014

Ahhh, ok.

@Tillerino
Copy link
Owner Author

That doesn't mean I won't accept pull requests.

On Thu, Oct 16, 2014 at 4:59 PM, otoed1 [email protected] wrote:

Ahhh, ok.


Reply to this email directly or view it on GitHub
#11 (comment)
.

@Tillerino
Copy link
Owner Author

There might be a simple way to tune down the difficulty of recommendations.
This was implemented in c30aac7 and is available for Patreons now for testing.
Not sure if anybody ever wanted to make recommendations harder 🤔, so if this works, it might be the solution to this issue

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

3 participants