Survival Mode Beta Test Brackets: A Framework for Testing
Survival Mode is one of the most exciting additions we’ve seen in Splinterlands in a long time — but it’s also clear the current structure isn’t working for the majority of players. If we want this format to succeed, we need to test something bold.
This post outlines a bracket-based Survival Mode aimed at creating meaningful competition for players across every level, while also giving the DAO a testing ground for new mechanics.
And to be clear: NONE of these reward payout numbers have been put through anything other than my own brain. This is a framework for discussion, not a finished product.
[IMAGE: https://files.peakd.com/file/peakd-hive/bjangles/23wC9v3PEYRBp4yXkft9QMWomPHnmne9isA6xLrjUSaPro6wUuHccU6Hg2dJsBo8izHW8.png]
The Proposed Brackets
- A total pool of 850,000 SPS per season for 3 months (~6 seasons) - set to serve as a Beta Test before making a DAO decision re: long-term parameters
- Reduced pool for the Beta Test Bracket test; intended to bump to 1.5m per season (ie., in-line with Modern & Wild Ranked) with identical percentage allocations per bracket once approved for long-term
- 10 brackets ranging from Novice to Unlimited
- Bot restrictions based on bracket
- Summoner caps, card level floors & unique mechanics for specific brackets
- Cooldowns across brackets are adjusted to reflect archon caps:
- See Splinterlands’ original post for Cooldown details: https://peakd.com/hive-13323/@splinterlands/introducing-splinterlands-survival-mode
- Novice: 10% of max cooldowns
- Silver: 25% of max cooldowns
- Gold: 50% of max cooldowns
This mix is designed to cover the bulk of player type.
[IMAGE: https://files.peakd.com/file/peakd-hive/bjangles/23wfjor3kjB5o9gTps23mS2FfGxd2shdfxgDGDFCXwAyvV3GCvdqDEgJTA4G8tcVorGXr.png]
[IMAGE: https://files.peakd.com/file/peakd-hive/bjangles/23wMSHvoqzhk4tKUEdVLqzgTL5ZqHGkacuN9jePryibwnJPjGH61FCj1bRXHXzLvobXey.png]
The Fine Print
- Liquidity Bots: E = Easy; A = Average; H = Hard
- You play within the bracket you’ve selected for the entirety of a season
EDIT: August 6
% Based Rewards: The Smarter Approach
I think the real answer actually lies in % based rewards. Instead of assigning fixed SPS payouts to each bracket, we treat the entire Survival Mode reward pool (say, 1M SPS) as a single pool shared across all brackets.
Here’s how it works:
Each bracket’s share of rewards would be determined by its total Bracket CP (Collection Power) relative to the Total Survival CP across all brackets.
Rewards would then be distributed retroactively within each bracket based on the same % formula, ensuring that payouts dynamically match actual participation and competition.
By “retroactively,” I mean that a player’s rshares (or another similar metric) would still be tracked for each win during the season as usual. Then, at the end of the season, the total reward pool for each bracket would be finalized, and each point or rshare would be assigned an exact value. Rewards would then be paid out in a single clean settlement.
For example:
Total Survival CP across all brackets: 500M CP
Bracket 8 CP: 100M CP (20% of total)
Bracket 8 would therefore receive 20% of the 1M SPS pool = 200K SPS to be distributed among its participants.
This model does a few things really well:
-
Eliminates arbitrary bracket payouts – rewards scale automatically with participation.
-
Encourages healthy competition – players know their share is tied to real engagement.
-
Removes the guesswork – the DAO and community won’t need to micromanage reward pool allocations every season.
In short, this adjusted rewards structure creates a self-correcting rewards system that grows or contracts naturally with player behavior—without the need for constant manual adjustments.
[IMAGE: https://files.peakd.com/file/peakd-hive/bjangles/23wgBZ5vVMMf1iM6fbBPvGDfrowG25iXBdFPAKaiy1HoTsmBStVT267y7Sfer3MgHpyCV.png]
Next Steps
I’d love to hear the community’s thoughts:
- Are these bracket definitions fair?
- Are the payouts balanced?
- Which bracket would you play?
- If this were an official DAO proposal, where would you stand?
If we get this right, Survival Mode could evolve from a niche experiment into one of the most important pillars of Splinterlands.
[IMAGE: https://files.peakd.com/file/peakd-hive/bjangles/23yTb75DaKfENumCQKdfKuRLth3vrKJAKWoWpA35jr2xCz3YhRePpiJSzRem9SAKZ3VJQ.png]
Until next time
@oaaguy | Aug. 5, 2025, 8:28 p.m. | Votes: 2 | [
VOTE ]
Hey @bjangles - great write up! I love the bracket approach with modified cooldowns based on level limits!
I have to agree with @keeegs here - too many brackets is no good for a trial...
- it will be overwhelming to decide what you want to lock into
- match liquidity will be lopsided;
- some pools will deplete much faster
I expect it will be a confusing and frustrating experience for many.
TDLR: Cut to 3 manual leagues and 1 botted league.
- Manual only Silver
- Manual only Gold
- Manual only Diamond/Champ
- War of the Machines mode - all card levels allowed.
Read below for my reasons...
I'd much rather ask: what does Survival mode solve?
First, Survival Mode is a lot of fun as a manual player in a high stakes game.
It allows players to think creatively in a way that ranked play doesn't.
I don't think we need a novice or bronze league version of survival. The costs of novice and bronze and silver decks are still fairly close, and we have Frontier mode specifically designed to guide players up to Silver level decks.
Survival mode is designed for players with lots of cards, but we still want to encourage card scarcity through combining.
By the same token, having liquidity bots with ghost cards in manual mode defeats the purpose of survival, and I'm 100% convinced that the bot operators aren't able to "tune the difficulty" consistently, and it creates too much work to keep adjusting them.
I'd much rather modify the matchmaking to address the known issues (see my earlier post with specific adjustments to the matchmaking algorithm)
Secondly, I do think there is benefit to a bot-friendly survival mode.
This provides utility for lots of excess unused cards, allowing the deeply-invested an opportunity to make use of their extra copies of max cards.
The whole point is we want to use bots in this mode to chew up cards. No one is going to say "I'm having so much fun botting my 10,000 Level 1 Kelya decks". Plus it will create too many matches for the servers if bots can play endless level 1 cards with 1/10 of the cooldown.
So just one botted survival league - and it's no league limits, no holds barred....
I'm saying we literally call it bot mode. If it wasn't copyrighted, I'd call it SkyNet.
In Summary: Cut to 3 manual leagues and 1 botted league.
- Manual only Silver
- Manual only Gold
- Manual only Diamond/Champ
- War of the Machines mode - all card levels allowed.
As for Playing in Multiple Brackets?
I think players probably should only be able to pick one manual league. Allowing play in multiple leagues will make understanding cooldowns more complicated ("Is the cooldown right? Which league did I lose my Tofu?").
Also for overall player satisfaction, it makes players feel weird when top accounts dominate low level leaderboards, etc.
But I see no problem with allowing players to play survival manually and in bot mode - My thinking though is that cooldowns from one mode would also be on cooldown in the other bracket.
Hey, really appreciate the thoughtful breakdown—and I do think a lot of your points make sense in isolation. That said, the goal of this whole idea, imo, is to test a system that invites more player types to actually show up & participate.
We’re not launching a final product here—we’re running a sandbox.
So while I get the argument for trimming things down to 3–4 brackets, I’d argue the reverse:
If we don’t test more, we won’t know which player segments are showing up, which brackets are working & where we’re missing traction.
A few quick clarifications:
Liquidity across brackets is something to watch, but that’s precisely why this format matters—let the data show us where things are sustainable.
I agree the system needs to be easy to understand—but I’d argue locking people into one bracket per season actually simplifies that, not complicates it.
Bots are here because we can’t ignore them—this mode can either absorb card supply or be gated for manual diehards. This test explores both angles side by side.
If we limit the menu too early, we end up baking in assumptions before we’ve seen actual results. The whole point here is player-driven evolution—the brackets with poor turnout, lopsided cooldowns, or confusing UX? They’ll be the first to go. But let’s start with options, and trim back from real feedback.
Again, appreciate the feedback—and I think once we’re through a full few seasons of data, this’ll all be a lot clearer.
