holoz0r's proposal to change the way the leaderboard and end of season rewards work

in Splinterlands9 months ago

image.png

The above screenshot is a quick mock up of the proposal. Imagine seeing a leaderboard for each league of ranked play in Splinterlands.

Click on Novice, you get the leading novice players, as well as their positions on the Novice league ladder. Click on bronze, gold, or anything else, and you get exactly the same.

A problem with this, however is - how do we deal with players who graduate from one league and find themselves instantly at the bottom of another?

By scaling the rewards in an intriguing manner, and making the thrill of coming last in gold just as good as coming first in silver. To complete a redesign of these mechanics, we'd have to first look at the current reward mechanics.

Warning: This is a long post. I do a lot of analytics, forecasting, and reporting in my day job. I'm trying to keep this as simple as possible, but it may be hard to follow this post if you don't understand some basic analytical principles.

Current Leagues

image.png

Each tier, and each league has a pretty nice bump in rewards. What's the real difference though?

image.png

What is a reward chest worth?

I took a sample of 125 reward chests from the PeakMonsters block explorer, or twenty player's reward chests. Someone else could automate this using some SQL, or blockchain data polling / API - I'm not that smart. The formula used to obtain the average DEC value of a single reward chest here is:

( Card Burn Value + Liquid Dec + Legendary Potions (80DEC value assigned) + Alchemy Potions (80 DEC value assigned) DIVIDED BY NUMBER OF REWARD CHESTS)

Using this equation and the below table, given my small sample size, the average DEC value of a Reward Chest is 134.08 DEC

image.png

We'll come back to these numbers. First, let's take a look at the current rewards for the highest tiers in the game...

Ladder Rewards

image.png

This means, each season, the following total number of booster packs are distributed to the overall top 50:

985 packs, or 1,970,000 DEC equivalent

In order to redistribute this pool of dedicated end of season prizes, we need to look at the distribution of players. Are there >50 players who finish in any given league? (I don't have the data, but probably) - if someone wants to generate a CSV, I can pivot that and do analysis, but if we reward the top 50 in each league, we would need 50 multiplied by 15 prize tiers.

That means we need 750 leader board rewards.

That's not a small number of "leader rewards". Dividing 985 packs into 750 people means there's not a lot of rewards available. But, dividing 1,970,000 DEC means there's a substantial reward for players. This is the approach I'm going to take from a value perspective.

Champion Leagues should get the highest awards

These players are the top end of the competition - and this is where the most powerful cards reside. These prizes should force players to be inspirational and want to join the champion leagues.

How do we then split the reward pool of 1,975,000 DEC fairly? I'm going to follow a typical split of the percentage of the reward pool for any given tournament set up by @splinterlands

image.png

This is actually perfect, as there are 16 tiers in my league tables above, so I'll remove the Tier 0 (the 16th, novice one) from the calculation...

image.png

Now if you follow what I'm doing here, the next steps should make sense, and you can skip this paragraphic. If you come first in a tournament you'll get about 15% of the reward pool for that tournament. Come second, its closer to 11%. Third, 8.37%, and so on down the ladder.

The other thing you'll notice is that there are essentially 5 tiers of prizes. For the end of the season rewards, however, there's six. Not too hard to divide this up.

Slicing up the reward pool for the end of season, if it was all average DEC value

Assumptions

  • 1,970,000 DEC reward pool budget
  • 15 prize winning leagues
  • Rewards to top 50 in each league

Table 1: Splitting the Pool Among the Leagues

image.png

Then, we split the pool even further for each level within the leagues.

Table 2


Splitting the rewards further

image.png

Each level within each league (and each jump - sees a linear increase in rewards. The rewards are more distributed within the player base. This only factors in end of season booster pack values, and not the DEC value of the reward chests granted by entering that league.

We then split the DEC pool for the TOP 50 in such a way that the 50th placed person in Champion 1, gets more reward than the first person placed in champion 2. I will not bother building charts for this, given the length of this post is also absurdly long. \

Table 3


Don't forget about the DEC value of end of season reward chests, using the average DEC per reward chest that I calculated above.

image.png

You'll see that the rewards scale in the same way as I propose LADDER rewards could be distributed...

Imagine ascending the top of silver. Pretty good feeling, right?

image.png

Then you're at the bottom of gold, and even though you've locked in your reward chests for getting to that league - you want to keep pushing, so even if you lose a few matches, you run the pleasure of trying to crack into the top 50 of that league level to get additional rewards at the end of the season.

This way, we have two schools of players

  1. Those that want to get as high on the ladder as possible.
  2. Those that just want to get to their comfortable league and grind those end of season reward chests.

This is my idea, and my suggestion and I believe it will get more people playing

This mirrors a lot of F2P mobile games and their leader board categories and reward tiers for free shit at the end of promotional or game periods.

You can even add a custom skin, or custom badge that the player earns for achieving a certain rank in a certain league in a certain season. I mean, what are Champion Points for?

This is easy to implement in code, as its a few if statements and a different end of season loop mechanism.

Now, I'm tagging a few @splinterlands team members to see what they think, and I hope they see this post!

@aggroed @yabapmatt @cryptomancer @nateaguila

Thank you all for your time

Sort:  
 9 months ago 

Great work. Putting figures on these concepts makes a huge difference.

Cheers, I do what I can, with what I have!

I still think we need separate leagues and your league is your choice (assuming you qualify for the league) and ranking system gets adapted to better indicate skill.

I think people may even play more and buy more because they may have an account in each league. Or they have one account that they try in gold and silver. But of course they cant get the rewards from both leagues but they can see how they match up.

Each league could have 3-5 levels of play giving it a continued feel of improvement and a bit of the grind.
And so top level people usually get to play other top level players assuming there is a match up at that moment but you could end up against lower players.

And then none of this manipulation of where you tend to get more points on average than you loose. Go back to a regular elo that attempts to properly rank skill. A bad player or bad bot would be rated low. I feel like the system right now makes even a terrible silver player able to hit diamond. Therefore our system plainly does not have an elo that is interested in ranking skill. But we would do well to have a ranking system that ranks skill for each group of users meaning those who just started with a few cards. Top players who can spend low hundreds. Those who spend high hundreds then those who spend thousands and then the pro/top levels.

In fact @yabapmatt and @aggroed should consider an invite only pro league where top next level players get invited and the lowest pro level players may have to move down to diamond if they are the lowest players in the pro/master league.

However your ideas for varied rewards based on levels works well. But a user should make a real decision... do I try to play diamond league where rewards are much higher but competition is much more difficult. Maybe get a top 50 rank... or go for Gold league and try to get top 10 ranking.

Insert 100 emoji here. Thank you for expanding on my idea.

You definitely put a lot of thought (and work) in this. I like the idea... a lot!

Thank you, it took a few hours to collate my thoughts (which I've had for many, many days) - and to build out the tables in Excel to see if they made sense. I'm glad that they did.

Clearly a lot of work went into this post. I agree with what your suggesting, not sure how higher ranked players will feel. But I reckon your right about making it more appealing to new players.

Thanks mate. Higher ranked players should welcome this, as it would lead to player growth longer term (and by extension) - increase the value of their cards as more people aspired to reach champion league.

Plenty of data, great article!

I have proposed to unify leagues, i.e. Champion 1,2 & 3 just into single Champion for the time being, due to a limited number of players with maxed cards and to reduce waiting times. But I have no precise data to optimize the number of teams in each league. Apart from Champion and maybe Diamond, it seems there are enough players to run separate leagues in each colour, i.e. Bronze I, II, III and the same with Silver.

Another potential source of rewards for players finishing high in the league are season end chests. For example, Gold II reward can be reduced from 26 to 24 chests; if there would be 400 players in the Gold II, we get 800 chests to extra reward top N players.

I like this idea as well - combine both our ideas, and I think we have a winner. :)

Great idea and gives more incentives. Might cause some just to try and hover around the top of a league though, rather than push into the next league.

Makes sense to create more incentives!

The bottom of a higher league should have a higher overall reward, based on reward chest value and quantity granted at the end of the season.

The breakdown of rewards is great but I see some problems. Say I am at the top of Diamond II with only an hour left for the season. Why would I continue playing and risk winning which place me at the bottom of Diamond I and get no rewards?

Two ways to solve this.

Rewards higher at the bottom of a league than the top of the previous one.

Combine each division of the league into one.

So instead of Diamond 1, 2 & 3, it is just Diamond.

I like the idea of "smoothening" reward steps a lot.

There are details that needs to be solved though

  • it would be beneficial to reach for example lowest gold and than intentionally drop to silver and play until you hit exactly max rating for this league.
  • if I remember correctly there was less than 50 players in Champion I
  • value of potions varies with bulk and guild discounts

Btw is this some kind of new player discount? Legendary potion cost me 40 DEC instead of 80. (Gold potion costs me 50)
Splinterlands_Leg_Potions.jpg

gold and legendary potions are priced differently based on their potential value. 40 dec and 50 dec is correct.

Some interesting ideas here. I love how you illustrated everything! But I honestly think the proposals by hotbit and rentmoney with fixed league for the whole season has more potential.

There's always multiple solutions to the same issue, and the same problem. They can be combined, or we can learn that an idea isn't good - that's the benefit of being able to debate things and discuss them in depth. :)

Great work and I like your idea!
I hope that the Devs will soon work on a better league and Reward system.

Thank you. Time will tell, and we'll see what is in the pipeline

The most annoying thing in the game are those players that play in lower leagues to earn more DEC, you find at the middle of the season, and almost with 3 days left, in gold league gf maxed decks. That should be changed.

 9 months ago Reveal Comment