Introduction
Since the end of 2023 year, Gwent: The Witcher Card Game is no longer actively developed by CDProjekt Red . Therefore no new cards are released and existing ones are not reworked. Nevertheless, to not leave the players with a completely stale, terminal meta, monthly Gwent Balance Council (BC) has been released along.
BC lets Gwent community change power and provision of cards via democratic voting. In total 40 cards are changed each month, 10x +1prov / 10x -1 prov / 10x +1 power / 10x -1 power.
A purely democratic system was established and we can observe its evolution and modus operandi month by month.
In this article we would have a general look on Balance Council impact in Gwent from October 2023 up to April 2025 and see what we can learn, what can possibly be improved in BC design and what analogies with real life democracy appeared.
Gwent Balance Council Beginnings
The first Gwent Balance Council voting happened in October 2023, resulting in November patch. Prior BC rules used back then were:
- 15 Changes per bracket
- Eligibilty:
- At least Prestige 1 (to discourage creating new accounts just for balance council)
- Be in Pro Rank or at least 25 ranked wins in the current season
(more rules from this time could be found in the first instance official Gwent Balance Council FAQ).
As Gwent players can see here, buff brackets were very focused on instantly impactful changes, which is especially visible in the power buff bracket. That being said, these impactful changes were mostly “new stuff”, barely present on Gwent ladder. Kaedweni Revenants, Cleaver, Crowmother, Ruehin, The Flying Redanian or Phoenix were so unplayed back then, that a huge effective power buff to them seemed rational.
The impact of balance council advice from influencers was overall lower then. As a result, many leader abilities were close to the top of -1 provision bracket. That’s a simple reaction: when we find a deck too strong and don’t want to go deep into which cards are problematic there, just nerf leader.
Leader nerfs are also focusing better than card nerfs. You can nerf a deck in many ways via cards, but only one way via leader.
If you look at nerf brackets, 50% of them is grey. That’s Nilfgaard (NG) faction – historically the most played, loved and hated one. The overnerf of NG resulted from strong emotions in the community rather than balance considerations. NG was in good shape back then and objectively deserved some nerfs, but at least in two cases hate was addressed very wrong: Vilgefortz and Alba Cavalry got reverted for good in the next patch. Both cards are of control type and were viewed as “fun police”. Two other cards got into “oscillating” state between stronger and weaker value: Slave Driver and Nauzicaa Sergeant. The latter one changes every month to this day with not a single patch ommited!
Next to NG overnerf, the 2nd thing which stood out was the power nerf to Reaver Hunters. After 2=>1 power nerf this card becomes totally unplayable. This change resulted from hate towards Reavers archetype. Luckily Hunters got reverted in the next patch and not nerfed this way anymore later.
This first balance patch was testing the grounds. Based on the results, CDPR conclusions were that too many changes are forced per bracket, the tail of lower voted changes is questionable and too much chaos is introduced. The next balance patch which preceeded Gwent Masters 5 Finals was limited to 5 changes per bracket only – making sure nothing unreasonable is done by the community before the last farewell.
The final form of monthly Gwent Balance Council which obeys till this day is as follows (official faq #2):
- 10 Changes per bracket
- Eligibilty:
- At least Prestige 1 (to discourage creating new accounts just for balance council)
- Be in Pro Rank or at least 50 ranked wins in the current season
So 15=>10 changes per bracket and 25=>50 wins adjustments were made by CDPR with respect to the first BC. A bit more stability was provided with these changes and the game was left alone to evolve…
What Balance Council Has Done So Far?
Let’s picture the power of decks/cards before Gwent Balance Council. First, let’s group them by power into following groups:
- U = ‘unseen’ = cards which weren’t played because of being underpowered, undiscovered or lacking good synergies
- T3 = Tier 3 = cards which see play in established decks, which however are not the best for climbing ladder for average player.
- T2 = Tier 2 = established decks decent enough for climbing for average player
- T1 = Tier 1 = top picks to climb the ladder in a given month; yield the best results on average
Now let’s put these groups on a power axis, from weakest to strongest, including the population of each group.
Before BC
The situation before BC would be as follows:
===============U=============== | =====T3===== | ===T2===| ==T1==
What balance guided council is supposed to do is to reduce the U by growing T3 and T2. There are two main approaches:
- Weakest First (WF) | Start from the farthest left (weakest cards). Such cards would be moved inside the U group and join T3 only after a couple of buff cycles.
- Weakest Last (WL) | WF isn’t wrong, but very slow. Better to start with cards close to T3. This would make meta more varied faster, while the weakest cards can be addressed last.
The argument between WF and WL supporters is everlasting. Most balance coalitions prefer WL over WF.
After 19 BCs
============U============ | =======T3======= | =====T2=====| T1
T1 shrinked. Best cards are less outstanding than they used to be – for example King Demavend, Svalblod or Melitele took 4 nerfs.
T2 has grown at expense of all other groups. The meta is more varied, more decks can be used for climbing.
T3 has net grown at expense of U. This was supposed to be the main process ongoing in BC, but doesn’t happen as rapidly. What is the cause?
Nerfs
So far we thought in one way: bringing the U up. But there are as many nerfs as buffs to be done. The midpoint of the power level is probably somewhere in Tier 3. Bringing the power of cards from the right to this point should make it easier for U population to jump into higher tiers relative to them.
In reality though… here is the major lesson to take from Gwent Balance Council as a whole: human psychology matters, has to be predicted and accounted for. Let’s look at possible nerfs:
- Tier 1 => high Tier 2 | Likely accepted by the community
- Tier 1 => low Tier 2 | T1 deck was likely popular; ~80% chance of revert to at least one card
- Tier 1 => Tier 3 | THEY KILLED MY DECK!!!! COMPLETELY UNPLAYABLE!!!! (in good hands still much above 50% winrate, but ALL DECKS SHOULD BE 50%+ WINRATE FOR EVERYONE) 100% revert
- Tier 2 => Tier 3 | WHY NOT NERFING TIER 1? (because if there are 2 decks in Tier1 and if they get 10 nerfs, then look up 3., BUT WHO CARES) WHY THIS DECK, THERE ARE DOZENS IN TIER 2? WAS THIS DECK ANY PROBLEMATIC? 100% revert
As we can see there is a huge psychological barrier, human factor which inhibits moving to the left. For the sake of justice, one has to add here that also not all cards can be conveniently nerfed, especially bronzes which are often in-between unplayed/strong state.
Nerf Barrier (NB) is the main, but not the only reason why out of 761 changes made so far in BC, 218 were net nullifying each other (~30%).
NB doesn’t exist that much for “sanity” or in other words “hate” nerfs. These are ones not based on balance considerations, but gameplay experience. Provision nerfs to Defenders or nerfs to Cultists/Mill are good examples here. Note that most of those nerfs easily sticked. It is not a matter of people hating those cards, because number of haters means nothing for buffs (reverts). Simply they are not as important and popular compared with Renfri or Jan Calveit to grab enough advocates from the silent majority.
Obviously, “hate” nerfs do not address balance issues and shouldn’t be made more often than necessary. We have to not only live with the Nerf Barrier, but include it in every real balance strategy.
A good example here is the placeholders discussion. Placeholders are used to fill the bracket to prevent a real nerf from happening. Obviously we don’t love the idea, it just slacks moving T2 to the left, right? Absolute nonsense. And powercreep! Because there are more buffs than nerfs, there must be powercreep!
What if we respect the NB here though? Extra nerf done instead of using placeholder indeed moves us to the left for the next patch. What if this one nerf was too much though and a revert comes? We not only wasted a vote for the nerf, but indirectly blocked a buff slot for the next council. We can only regret not picking a placeholder then – in the long run would have wasted one slot less!
Sometimes a real nerf would be better, but sometimes the placeholder… Similar discussion is made about leader cap buffs, which use a nerf slot.
Here I don’t see any reason to treat leader cap different to a card. There is nothing wrong about buffing Tier “U” leader ability. Taking Nerf Barrier into account, buffing Tier 3 leader may sometimes also be considerable. While buffing Tier 2 / Tier 1 is wrong, just like for normal cards. The effect of a wrong buff here is doubled, because revert requires using a buff slot, which is more effective (because of NB).
If the explanation above is sufficient for you, then you have never been on the Gwent Reddit. Apparently I’ve made a typical error here (shortsighted, not educated enough) because using nerf slot for a buff = 21 BUFFS vs 19 NERFS = BREAKING THE SYMMETRY OF THE UNIVERSE = POWERCREEP = INFLATION. In fact buffing a “U” leader ability may impact damage to the symmetry equivalent to a PRINCE STENNIS GETTING INTO BC TRAIN WIHTOUT TICKET (POLICE!!!!).
Historically there are examples of “good” and “bad” leader ability buffs, with the latter being almost exclusively T2 buffs, coming from “impact hunting” which will be discussed below.
Impact Hunting
The higher tier is the deck the more it impacts the meta. All buffs which do not result in immediate shift amongst Tier1/High Tier 2 do not change the state of the ladder much.
Most work here should be done by nerfs bringing top decks down, but as we’ve seen, Nerf Barrier warps this process. Therefore some balance coalitions lose patience, or even pick instant top meta impact as a regular strategy. Meta shifts are fun, buffing a card from U to Tier3 to still be unplayed in mainstream is no fun.
Such approach is in obvious contradiction with the general picture. Moreover, cards buffed this way become natural points of focus for next season nerfs. That’s another high contribution to 30% of wasted slots in Gwent BC history.
In the Nerf Barrier reality picking a bit more impactful buffs is reasonable, but one has to make sure 1) the buff doesn’t end up as Tier1, 2) the real power of buffed stuff wasn’t already Tier 2.
Successes And Failures Of Gwent BC Design
Successes
- Homeostasis – at the start of Gwent Balance Council the community could have been afraid of “selfish Balance Coalitions” which will align votes to force buffs to stuff they love to Tier0 and nerf stuff they hate to Tier U. It just doesn’t happen. Even after most visible “impact hunting”, in the next season the cards of interest are reverted by the silent majority. Such reverts usually have more votes than any coalitions suggestions. That’s the main function of democracy – fixing changes which are against the interest of the majority.
- More Varied Meta – thanks to balance council, more cards see play, more decks and ideas are considerable. Of course, due to Nerf Barrier the popular decks from times before BC still have to be high in the meta, but they are accompanied by more new alternatives.
- More Playable Cards – if we want to go off-meta, then we can make use of more playable cards than before BC
- Reducing Binarity – Gwent Balance Council so far has done a good job at tuning down cards/decks which are binary and bad for ladder experience (Defenders…). The game became not only more balanced, but also in each match players have more agenda.
- Keeping Game Alive – Gwent is still interesting enough to attract new players or make some veterans like Pajabol or Demarcation return. Also Balance Council discussion is what makes community more involved in the game overall.
Failures
- Lovers/Haters Discoupling – number of lovers matters for buffs, number of haters matters for nerfs. These groups do not nullify each other – some cards like Nauzicaa Sergeant, Slave Driver, Renfri or Lord Riptide can have more lovers and haters than 99% of other cards at the same time. As a result we often walk in circles and waste 30% of votes by changing same cards many times. It would be perfect to make both groups involved in each voting, so that a consensus can be achieved. One solution I’ve seen suggested in the community is to make the voting two-stage. First stage ~Top20 cards for each bracket are determined in the popular vote. Second stage from the pool of known cards the community votes (+) for changes they would like to accept and (-) for ones to be stopped. Top10 net results from all brackets are the final result.
A possible downside can be worse “Homeostasis”. As a sidenote, I’d like to see sth similar in the current presidential elections in Poland. Instead of putting just one cross next to one candidate, how about having (-/=/+) options for each one of them? We lose so much information and adjust our votes to poll results so much right now. Also we can’t tell politicians how much we like them in a measurable way
- Nerf Barrier – as discussed in the “Nerfs” section, in practice nerfs are far more troublesome than buffs. Beacuse there is far more non-controversial buffs to make than non-controversial nerfs, it would have been better if the number of buffs/nerfs could have been adjusted with respect to each other. Sadly at the design level programmers left only “changes per bracket” lever rather than separate lever for each bracket. For example if buffs vs nerfs proportion would have been 15 vs 5, we would have far less wasted votes right now and smaller “U” section. Of course in the very long term when balance reaches equilibrium, 10 vs 10 is better so that the game doesn’t drift in power. But we are very, very far from this state and likely wouldn’t reach it in Gwentfinity lifetime.
The main lesson is: leave more levers even if you don’t see any use for them rn. The analogy with real world democracy is also powerful. Big groups of interest just have to keep their privileges or will riot. Democratic government will make sure that the (financial) situation of these groups is at least no worse, even if it means increased national debt (setting the midpoint in Tier 2 rather than Tier 3) and a bad impact for the state as a whole. - One-Season-Wonders – community is no different to developers – if they want to do something stupid, they will do it and fix later. Nothing can stop popular, but bad ideas to be one season wonders. The most vivid example here is Seagull buff to 2 power in Feb 2025 patch. The two-stage voting idea shown in “L/H discoupling” is one way to address this failure.
- Community Bubbles – Gwent community is split into different bubbles. Especially language differences are important here: Russia, China, Poland, US, Ukraine, Germany, Japan… – many countries have impact on the final result. In such a situation the exchange of information and especially going down to the roots (a single voter, often embedded only in a local community) is very hard. Good point made about Gwent BC in China may never diffuse to Russia or EU and reverse.
To the best of my knowledge Gwent Reddit is the only global BC idea market of significance and obviously it is imperfect.Maybe if there was a multi-language forum dedicated only to Gwent Balance Council, advertised and clickable via game client the situation would have been better. Anyway, such stuff would have required really good translator and moderation.
The bubbles are also essential to real life democracy. I don’t even mean here the political parties. The step of going from local (bubble) to global popularity is crucial for the success in Democratic elections.
- Expertise Gap – there is no question that some players understand the game much better than the others, even if both are on Pro Ladder. Pajabol’s vote weighs as much as hello123 vote. No offense, but the percentage of players who understand the game on Gwent Pro Ladder is much lower than 50% and lower than 10% too. As a result, the balance council votes reflect the casual experience more than the real state of the balance between strong decks. Stuff from the very top of ladder/tournament meta can be unaddressed for years unless popular. Top players are forced to publish own suggestions in order to reduce the gap and address some unpopular stuff.
I think that BC would have worked a bit better if the votes were somehow weighted by Gwent Pro Ladder position. Such weights would not only make up a bit for the expertise gap, but spice up the ladder rivalry, especially if the ladder winner had a significant weight bonus. Finally, this way we reduce a bit the impact of multi-accounts made just to cast more votes in BC.
Of course, there is also strong opposition to such kind of systems. That’s another powerful Democracy analogy: my vote must always have the same weight as the vote of the others. Discussion here is pointless – we are in the “fundamental laws” region. One can only observe that we need a license to drive a car, but need no license to elect people who make licenses.
Closure
To sum up, Gwent Balance Council is overall a success and I can recommend a similar solution (maybe including improvement suggestions from this article) to every game going into dormant state. It just works!
I planned to fill this article with more analogies to real life democracy (nerfs = taxation = no good taxes… and so on), but the amount of text grew too fast. Gwent Balance Council can be studied from different angles. I didn’t touch much the hot topic of balance coalitions and gaining support either. I just didn’t find it that important compared with the overview of how BC works for… Gwent.
As you can guess from % of wasted votes, I have done some stats work on BC results and may publish some data soon. Let me know what kind of stats/graphs would you like to see.
Cheers,
lerio2