Jump to content
Stray Fawn Community
  • 1

Sumo Weight Division


Lulzer

Post

The problem with sumo is that often the dumb, big drone will win over the small, clever design. There's no incentive to build a compact, well-thought out drone using fewer parts because mass is a huge factor in a drone's ability to remain inside the ring. Constraining a drone's participation to a 'weight' class (number of parts) would open the field for a greater variety of drone designs.

EDIT: A minimum number of parts would also prevent things like this. What exactly is the criteria for the drones that are drawn as opponents in a tournament? What about those selected to 'guard' a planet in exploration mode?

Captura de pantalla 2018-10-13 16.38.20.png

Link to comment
Share on other sites

15 replies to this post

Recommended Posts

  • 0

Drones with at least 1 point (1 win) get into the tournament pool. So this poor empty drone brain was just lucky in the first fight.
The drones that are placed in the ring are randomly taken from the leaderboard according to the difficulty level of the arena:

1 Δ = lowest 20% 

2 Δ = 20%-80%

3 Δ = Top 20%

Weight classes are a good idea. I'm also thinking of adding a higher difficulty pool where only drones with 10 wins can enter.

  • Like 1
Link to comment
Share on other sites

  • 0
18 minutes ago, Micha said:

Drones with at least 1 point (1 win) get into the tournament pool. So this poor empty drone brain was just lucky in the first fight.
The drones that are placed in the ring are randomly taken from the leaderboard according to the difficulty level of the arena:

1 Δ = lowest 20% 

2 Δ = 20%-80%

3 Δ = Top 20%

Weight classes are a good idea. I'm also thinking of adding a higher difficulty pool where only drones with 10 wins can enter.

Is the leader board public? Or can we at least see where out own drones rank? That would be nice.  What exactly is the formulation for leader boarding? As you know livetesting may impact this. 

Link to comment
Share on other sites

  • 0

The leaderboard is pretty simple, it's just the score you get (amount of wins) that's why we don't show it.

The main problem thats blocking us from switching to some better rating system is that Steam leaderboards don't allow to edit entries of other users.
As the multiplayer is not real-time we can only change the score for your own drone and not the score of the opponent.

Link to comment
Share on other sites

  • 0
12 minutes ago, Micha said:

The leaderboard is pretty simple, it's just the score you get (amount of wins) that's why we don't show it.

The main problem thats blocking us from switching to some better rating system is that Steam leaderboards don't allow to edit entries of other users.
As the multiplayer is not real-time we can only change the score for your own drone and not the score of the opponent.

I think I see. So.. Is it like, the number of wins in the tourney last completed? or is there an accumulation and averaging at all? Are there a bunch tied for first with 10 wins? 

Link to comment
Share on other sites

  • 0

Just the number of wins of your last tournament together with the drone.
I was thinking of using a seperate leaderboard where only drones with 10 wins enter automatically and everyone can try to battle them.
 

  • Like 1
Link to comment
Share on other sites

  • 0

Can you entirely reset the leaderboards?  That would seem fair enough, if you're implementing an entirely new system.

Alternately, you could use values that rank a thousand points for meager displays, and 10,000 for a fantastic showing.  Divide by 1000 to display the result ingame. (Instead of 956,789, display 95.6789.)  

You would dump nearly every steam ranking based on # of wins straight to the bottom until they played the game again, and established a ranking in the new system.

Disclaimer: I have no idea how the steam ranking systems work.

Link to comment
Share on other sites

  • 0
18 hours ago, Lurkily said:

Can you entirely reset the leaderboards?  That would seem fair enough, if you're implementing an entirely new system.

Alternately, you could use values that rank a thousand points for meager displays, and 10,000 for a fantastic showing.  Divide by 1000 to display the result ingame. (Instead of 956,789, display 95.6789.)  

You would dump nearly every steam ranking based on # of wins straight to the bottom until they played the game again, and established a ranking in the new system.

Disclaimer: I have no idea how the steam ranking systems work.

By what metric would you rank a "showing"? 

Link to comment
Share on other sites

  • 0
15 hours ago, Alpino_WILL_STEAL_ oats! said:

Actually the small clever drone almost always wins if it has a stratagem.

You mean if you build a small drone to beat a specific drone? Maybe. Unlikely. There's just no way to get a hundred-part spinner with basic center-seeking out of the arena with a thirty-parts or less drone. But you build whatever you like under thirty parts and I'll pit it against that dumb spinner you see in my OP (and that particular spinner is under sixty parts). Will post results.

Link to comment
Share on other sites

  • 0
5 hours ago, RAC said:

By what metric would you rank a "showing"? 

Take your pick of the suggestions hereabout.  Average of your last 50 matchs, with an unbroken record being 10K, and only a single win being 200.  The amount of time it takes you to win, with a last-second victory being a low score.  Maybe factor in the damage you deal.  

There are a number of ways to represent the metric.  The point is, if you change the FACTOR that scores are represented with, you can overshadow old scores with new results.

Link to comment
Share on other sites

  • 0

Will play devil's advocate.

- Average: Sounds like the strongest metric if you could get it to work. Problem is the limited pool of drones that will have played X number of matches. Part availability being what it is, I haven't fought with the same drone more than 10 or 15 times, and many of those were in single player where I keep a "strong" (dumb and big) drone to progress through the galaxy quickly. Plus, you run into a Theseus ship kind of thing pretty quickly. How many modifications are allowed on the same drone before you need to reset its score? If I swap out a single part, is it still the same drone? What if I change the inputs and outputs? This is handled right now in some way, How is the game tracking a drone that goes through modifications, big and small, over time?

- Time to Victory: Last second wins you argue are low-score because it comes down to a bit of luck as to where the drones are when the ring shrinks, right? Insta wins would also be low score because it implies you wrecked the other drone, maybe you dodged a blind pusher - that's 3 thrusters and a detector, not exactly high tech stuff. So the high score would fall somewhere in the middle, say between 20 to 45 seconds. That's not a bad metric, if a little arbitrary.

- Damage dealt: Entire matches go by without visible damage. Also, how would part destruction that is self-inflicted or because of jiggly physics contribute to the showing score? Question for the devs: in sumo mode, parts exploding happen because a single collision exceeds an HP threshold or is there cumulative damage applied to each part?

And yeah, change the factor enough and you make previous scores obsolete. You can just wipe them, too. It's early access after all.

Link to comment
Share on other sites

  • 0

If the drone hasn't played 20 matches, or whatever number, use the average of those they have played.  The ranking is quality, not quantity.

There are numerous suggestions for applying the ranking.  I'm not trying to tell them how they should rank - do it like you do now, if you like, just multiply scores.

This isn't about how matches should be ranked.  It's about how to alter your ranking system, AND re-rank already ranked players, without being able to alter the Steam leaderboards.

Link to comment
Share on other sites

  • 0
6 hours ago, Lurkily said:

Take your pick of the suggestions hereabout.  Average of your last 50 matchs, with an unbroken record being 10K, and only a single win being 200.  The amount of time it takes you to win, with a last-second victory being a low score.  Maybe factor in the damage you deal.  

There are a number of ways to represent the metric.  The point is, if you change the FACTOR that scores are represented with, you can overshadow old scores with new results.

Yeah, some averaging for sure. I think now its simply 1 steam account - 1 drone saved,  the drone that is saved =  whatever drone you last went through a tourney with - so long as it won at least once..  Thats my impression from Michas comments. So clearly there is room for making the system more .. robust. 

Link to comment
Share on other sites

  • 0

My impression was that it's a per-drone ranking, but I'm mostly worried that the leaderboards on Steam are inviolate; that makes them the barrier to making rankings better, unless we do some drastic silliness like this.

If we do use averages, I would start the player off with one symbolic loss in the average, so the leaderboards aren't crowded with one-match wins at 100%.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...