Let's Fix The Brawl System - Take 2

in Splinterlands6 months ago

Alright, so after a lot of discussions, getting feedback from a lot of stakeholders and also a major discussion in German with @schachoberhessen over Discord (thanks for that!), I decided to give it a second shot. My goal is to solve the issues the Brawl system is facing in a lasting way instead of only adding band aid solutions.

I'll try to explain everything as thoroughly as possible because as of now, it often seems like it's not exactly clear what a proposal would actually do and wouldn't do. I'll start by describing the current issues with Brawls and how I propose to fix that.


splinterlands.png

The Issues

  1. Many guilds are playing below the highest tier they could actually play in because they will earn more Crowns&Merits playing in a lower tier than playing in their highest possible tier.
  2. The top guilds earn over proportionally more Crowns then the rest of the pack, thus making it impossible for everybody else to catch up or even keep up with the top guilds.

The Cause

There are several causes coming together to foster these issues. First of all, though, it's important to understand why (the top) guilds are staying in a lower tier:

The most important resource you get from Brawling are Crowns, simply because you can only earn them from brawling. These are followed by merits, which can also be won in chests but still can't be bought. SPS is the least important factor in Brawls because if you want more SPS, well, you can just buy it or earn it through all the other options in the game. So guilds care for earning Crowns&Merits first and the fact that they could earn higher SPS in a higher tier isn't motivating them to move up.

Now, why are guilds earning more Crowns in lower tiers? There are 2 mechanics at work here:

  1. The amount of Crowns given out is calculated by the amount of battles played and only taken into account for players that won at least one match. So with 10 guilds in a Brawl there are obviously way more battles fought than with 5 or 6 guilds in a Brawl and thus there are way more crowns given out.
  2. The top 3 guilds in a Brawl get a bonus to their Crowns earned, making it especially interesting to stay in a tier you are dominating compared to competing in a tier that you are only placing 4th or 5th.

The effect of this can be seen for tier 5 right now. There are 7 guilds that could compete in tier 5. Playing in tier 4, they are almost guaranteed to go 1st or 2nd unless all top guilds end up in a Brawl together. Also, in tier 4, you usually have 8 to 10 guilds per Brawl, thus increasing the Crown pool compared to tier 5 significantly.

Unless these causes are addressed, the aforementioned issues can't be cured. Even if more guilds advance to tier 5, we would face the same issue again once there are 11 guilds ready for tier 5. This would result in a 5 guild and a 6 guild Brawl, again making everybody earn less Crowns than by just staying in tier 4.

The Solution

I'll do a TL;DR of what I would like to propose in the end but will first explain it in more detail here.

The first part should be a quick and easy fix and is part of what @schachoberhessen already proposed. Remove the top 3 bonus and distribute Crowns&Merits among all guilds according to their win rate only. This alone is not going to fix the issue of having not enough guilds in a Brawl, though.

So on top of that, the amount of Crowns&Merits should no longer be dictated by the amount of battles played in a Brawl. Instead, the amount should only be dictated by two factors:

  1. The Brawl tier
  2. The win percentage of each guild

For this to work, a no-show in any fray would be considered a loss and the win rate would thus be calculated as number of matches won divided by number of possible matches in that tier.

By making this change, it would always be worth it for a guild to advance to a higher level tier because they would be guaranteed to play for a larger Crown&Merit pot than by staying in a lower tier.

To further address the issue of the top guilds earning over proportionally more Crowns than the rest, the formula for Crown distribution is also to be changed. Instead of using a linear system we would move to a hyperbolic tangent formula which would look something like that:

The formula is made by @kalkulus which has been an awesome help in setting this solution up - thanks a ton man!

I'll first give you the formula and then explain it with an example:

crown_reward = tier_average x (1+tanh((winrate-50)/25))
(where tier_average is 16.25, 44, 76.5, 120.75, 187.5 for tier 1,2,3,4,5, respectively.)

The goal of this formula is three-fold:

  1. Keep the average crown payout where it is right now
  2. Move more of the crowns to the averagely performing guilds
  3. Still make every additional win matter

So how does the formula work? First, we take the average crown earnings per Tier which looks like this:

  • Tier 1: 16.25
  • Tier 2: 44
  • Tier 3: 76.5
  • Tier 4: 120.75
  • Tier 5: 187.5

So for tier 5, the expected crown reward would look like this:

187.5 x (1 + tanh((winrate-50)/25))

If you are like me and can't compute tanh() in your head I'll give you some examples of what that would look like:

0% winrate: 0 Crowns (this is actually not the result of the formula but the obvious value)
10% winrate: 14.69 Crowns
20% winrate: 31,19 Crowns
30% winrate: 62.99 Crowns
40% winrate: 116.26 Crowns
50% winrate: 187.5 Crowns
60% winrate: 258.74 Crowns
70% winrate: 312 Crowns
80% winrate: 343.81 Crowns
90% winrate: 360.31 Crowns
100% winrate: 368.26 Crowns

So as you can see it starts and ends slow but accelerates quickly between the 20 and 80 percent mark. This would help to bring Crown earnings closer together across the board while ensuring that guilds that won more would always also earn more crowns.

Lastly, the Arena building would lose its effect of boosting Crown earnings. In my opinion, this wouldn't be too harmful as a higher Arena would still allow you to compete in higher tiers which, with these changes, would always be beneficial anyway. This would leave arena level 10 open but I don't think this needs to be solved right away.

Right now, no guild has a level 10 arena and I also don't think any guild is planning to build one as it doesn't really make sense with the current incentives either. For these reasons, I would just change the Arena as described and make adjustments in later proposals if necessary or when somebody comes up with a good idea.

The TL;DR

Here is what I propose:

  • Remove the top 3 bonus
  • Base Crown payout solely on Brawl Tier and win percentage of each guild
  • Move to the hyperbolic tangent formula crown_reward = tier_average x (1+tanh((winrate-50)/25)) for reward calculation
  • Change Arena to only gate the Brawl Tier you have access to

Your Turn

And now I'd love to hear your opinion. Would these changes work for you? Is there anything that I didn't think of or do you have any questsions?
Please, feel free to leave a comment here or hit me up on Discord. If we can build some amount of consensus here I'd love to forward this as a proper pre-proposal.

Thank you all for reading and see you next time!

Sort:  

Awesome that you are tackling this issue. I think you fully understand the problem and love how much effort you've taken. I agree specifically about the top 3 bonus for crowns and merits (the SPS isn't my concern).

The math in your formula is over my head, but I trust that it works as you've explained.

Great job and I hope you and @schachoberhessen can get attention on this issue so that the brawl experience can continue to be a driving factor in the game!

Thanks a lot for your feedback.
To tell the truth, I would have never been able to come up with the formula without @kalkulus - he's the real MVP here.
But I know how to code, so once I have it, the rest is rather simple ;-)

This is well thought out and addresses the issues for guilds like mine, TeamPossible, choosing to stay in Tier 3 even though we should be Tier 4 or even possibly competitive in Tier 5.

Thanks for the thought and work. I'm curious what @yabapmatt and @cryptomancer think about the equation.

Thank you very much for your feedback.

I guess everybody is busy with the launch of Rebellion right now. Cryptomancer already told me he'll read it when he has the time in a few days 👍

I like this proposal, maybe you make arena lvl still make matter by having the average crown lvl multiplied by 0.8 for lower lvl in the tier and 1.2 with the higher lvl

Interesting idea thanks for the feedback. Something like that could work, let me tinker with the numbers a bit.

We actually discussed the second part that you are addressing here and that is where I got feedback from @cryptomancer, the dev that designed brawls to figure out the scope of the work. It would likely take a few months to get implemented if we vote on that issue now. That's why we are going for the pieces we can with the current proposal. We aren't ignoring the other issue, it's just not something that can be fixed at the current time, so we'll make a proposal for that later. In general I agree with all of your points.

Thanks for your feedback.
To me, the most important part is to fix it proper. I don't mind if we have to wait a bit until it gets implemented.
It's obviously not our highest priority, especially with what happened in recent months. Still, Brawling is an important part of the game to many and we're dealing with these issues for quite a while now.
I already contacted Cryptomancer to see where he stands on the topic and how much effort the implementation would actually be. So if there's some consensus for the change and implementing it eventually is feasible, I would like to bring it to a vote now so that this part is done. It can still be put on the road map for some time in 2024 afterwards of course :-)

I get it, we were going to actually open a second discussion channel for the counting empty frays issue. The problem with putting too many things in one proposal is there are multiple reasons for people to vote no. Like some people are already against the current proposal because of the boost reduction. We wouldn't want the empty fray proposal to fail just because the other part did. I think we're trying to get to the same place and I actually think this second part you want will be less contentious. We'll see how the current one goes, if it fails we can try to do a comprehensive proposal a bit later.

It's a delicate balance for sure. I've voted against the proposal precisely because it was too little in my opinion and I didn't want another band aid solution added on top. We have that tendency of throwing a first version out there and not coming back to iterate on it afterwards for months or years.
But I'll go ahead and change my vote there ;-)

Love this proposal. Would love to see this pass even if it cannot be implemented in one shot... the step by step method of band aid fixes and endless small proposals seems overly time consuming and excessive... can't we pass it all in one shot like this proposal suggests?
I have a feeling only those in the top three guilds who are benefitting would vote against this, and then only some as the author himself is in a top 3 guild and recognized the dire NEED to fix the broken system ASAP.

Thank you for your feedback Jimmy!
My plan is collect feedback for the rest of the week and post the actual proposal early next week (unless anybody comes up with any major points until then).

It's a great proposal. I hope all agree.

Thank you very much for your feedback :-)

This post has been supported by @Splinterboost with a 20% upvote! Delagate HP to Splinterboost to Earn Daily HIVE rewards for supporting the @Splinterlands community!

Delegate HP | Join Discord

I really like this proposal. Its excellent work done @khazrakh.

@clayboyn If parts a complex or taking long to develop, the proposal can still be voted on as a whole, but development can be split in several parts.

So easy things made first and others made later. No need to do all developments of a porposal at once in my opinion.

Thanks for your feedback and agreed, as mentioned in the conversation with clayboyn, I would prefer if a single proposal was to tackle the whole issue with several steps than voting on minor parts one step at a time. Obviously doesn't mean it all has to be implemented at once.

Sounds very good, I will support it.
About the Arena-Level as far as I understand it has influence on the whole pot of a brawl and not your personal earnings (or only indirect in a very small way) - cause it is random in what brawl you will be placed I really would like to see this taken away as you suggested.

Thank you very much for you feedback :-)

Thank you! I agree we lost a lot of time on this, the issues were apparent a long time ago already, but better late than never ;-)