While the card examples in this article are partially outdated, the theory behind this analysis is still correct. We will be updating this article to be in line with the latest Gwent patch soon.

Introduction: The Non-Binarity of Card Evaluation

The first thing we see when we look at a Gwent card are the numbers at the top left and bottom right. And if you look also at the ability, it is easy then to say, “This card is good” or “This card sucks.” The reality is that in any complex system, we cannot evaluate any card based on any intrinsic attributes alone.

If we look for a second at the Reward Book, whether a node is good includes not just the rewards it gives, but the nodes needed to access it and the nodes it gives access to. Every 1-point ore node would be the same, but some are more valuable than others because they lead to leaders or to a 3-point ore node. Similarly, cards that look equal in power or provisions may be very different based on their relationships with other cards and game mechanics. This begs the question: How do we evaluate cards?

We should not use a binary system, since this can never do justice to the interconnectedness and malleability of the card itself and the Gwent ecosystem. Every card, even a vanilla one, still functions better in certain archetypes than others; for instance, Shupe's Day Off can be a potent key card in an Emhyr var Emreis deck, but with duplicate cards in deck, it is worth 0 points. In this series, we will discuss some metrics by which to evaluate cards from a more holistic approach.

Main Factors for Evaluating Cards

No set of factors, and certainly not a binary yes or no, can quite capture the insanely complex web of almost 500 interconnected cards in Gwent. The spectra of attributes discussed in the following sections of this guide only capture a part of the system, and encompass many different aspects of card interactions. The non-binary system of factors we will discuss is not all-encompassing, but it edges closer to the “ideal” quantification of a card’s power.

        The Point Per Provision Valuation System

Point per provision (PPP) value is a concept that was introduced with Homecoming and the provision system. The purpose of this system is to act as a single, universal way of gauging the value of cards that can apply to any card in the game. An extremely low point 15-provision card will still get more points than a high point 4-provision card almost every time - but this doesn’t mean that the 4-provision card is less viable. So, we normalize the values by factoring provision cost into the equation, and now, we can see that 15 provisions getting 9 points (example only) is much worse than 4 provisions getting 6 points.

The accepted norm for “expected value” of a card is 1 more point than its provision cost. However, certain types of cards can achieve more or less than expected value in different scenarios, including synergistic cards, as one example. Seeing as the total value of a card cannot quite be quantified, we will only count raw points rather than denied points by Locks or other utility effects for the purposes of illustration. Now let's get into the first of the three evaluation parameters.

        Niche and Versatile cards

The first spectrum we will be looking at is the scale of niche and versatile cards. This is important to consider, as the position of any card on the spectrum of its flexibility in use cases or lack thereof affects the power level of that card and its competitive play rates. This can be partially attributed to the fact that inflexible cards tend to be matchup-dependent and sometimes draw-dependent but have extra value. More flexible cards are also needed across all archetypes to plan for all matchups and draw situations. Now, we will explore these two points on this spectrum individually and then tie the spectrum in to the big picture of card viability and balance on a high level.

Niche cards

Niche cards are useful in specific situations but underwhelming outside of them. These cards are by nature risk-reward-based since their PPP greatly varies depending on the scenario. Niche cards are often integral parts of many midrange and point-slam archetypes, and their gameplan involves creating the ideal situation for the niche cards to shine. One needs to strike a balance between these cards and less risky cards for successful deckbuilding. This will be discussed later in this article. To illustrate niche in a more specific, in-game context, we will look at two examples of niche cards, one which never was in the meta, and one which was, and try to find the reason behind this, to understand the characteristics of a good niche card.

First, we will look at Dandelion: Vainglory. Obviously niche, his ability is only useful with a Beast on the board. With its low chance of activating as many Beasts are faction-locked, its reward, although potentially quite big in terms of PPP, pays off too little to make it worth the risk.

The other card we will look at is the pre-hotfix Golden Froth, which had the same ability as current Golden Froth but at a cost of 9 provisions. While technically a niche card since its ability achieved high points in a specific scenario, it was easy to build around compared to the payoff it provided. Before the hotfix, Golden Froth was more of an ubiquitous point-slam card in many decks and achieved far above its PPP very often since it needed little setup. After the patch, it became more niche in that it now needs a bit more setup to achieve its PPP, and it does so less often. The 9-provision cost was so problematic mainly because it took away most possibility of consequence (in the form of lowered PPP) if the niche was not achieved.

We can conclude that the reason Golden Froth was more used as a niche card than Dandelion: Vainglory was not the magnitude of its payoff, but the consistency with which there would be any payoff at all, and the lack of punishment for it not paying off. This follows for other niche cards - one can evaluate them by thinking: How often will this card pay off, and what is the consequence if it doesn’t?

Versatile cards

In contrast, versatile cards are able to fit in in a vast number of situations, the most obvious of which being cards with dual-row abilities. Many vanilla cards tend to also be flexible or versatile, since their uses are not inherently confined to certain cases by the nature of their abilities. Boost and damage cards are some of the most versatile, as their targets include almost anything on the board. This can be seen as analogous to the fact that the closer to a computer system’s lowest level you get, the more freedom you have to work with hardware. Every deck needs adaptability in order to create more uniform performance across many matchups and possible card draw outcomes.

Now, looking at some in-game versatile cards, we can tell why decks need a balance of cards like these and niche cards that gain raw points. Versatility means wider utility in more cases, and cards that exemplify this usually have higher use rates in a competitive setting among other similar-PPP cards.

A great example of this are the two Scoia’tael Elves Dol Blathanna Archer and Dol Blathanna Bowman. They are on equal ground in nearly every way - provisions, faction, and tags - except for their abilities and power. Since we know that Dol Blathanna Archer is used more than Dol Blathanna Bowman, we can see that the cause is nothing but its ability, since all the other traits are the same.

The ability of Dol Blathanna Archer allows for productive use in more situations (killing own units or Spies, disaligning units, and many more) than Dol Blathanna Bowman. We can see with Dol Blathanna Archer, as well as other similarly versatile cards like Milaen, that in order to gain the versatile aspect of the card, we pay some points or provisions, and that the amount scales with the card’s power. A card like Dol Blathanna Archer will consistently get 5-value points + utility, whereas a card such as Xavier Lemmens has more of a risk in PPP and will not pay off every time. So in exchange for larger-scope consistency across games, versatile cards pay some points in the smaller scope (the current game).

Pre-hotfix Golden Froth was so consistent in gaining points that it almost had the characteristics of a versatile card - consistent value and broad usefulness. It had all the benefits of a niche card without the drawbacks and most of the benefits of a versatile card, with those drawbacks covered by its niche aspects. The provision increase made it more variable in its point output and more niche in that it became a risk-reward card like the other niche cards.

Knowing what classifies and identifies niche and versatile cards, we can investigate the impact that these qualifiers have on card balance. Having an understanding of balancing strategies can help explain balance change notes after they come out, or help predict and plan around impending balance. Niche cards are sometimes rebalanced because of their dominance within the niche (i.e. Golden Froth) or frequency of niche use (i.e. Xavier Lemmens). Often, these cards get the benefits of niche cards without the downsides.

When looking at cards and evaluating niche and versatility, one should take into account the best and worst matchups of niche cards and whether the possible payout in PPP is greater than the possible cost; and with more versatile cards, whether the versatility will be useful and consistent enough to justify the likely reduced PPP value.

        Point Floor and Point Ceiling

The next pair of factors worth considering are point floor and its inseparable partner point ceiling. These two are closely related to the other factor groups, in that a card’s position in those spectra manipulates its floor and ceiling. Point floor is the fewest number of points a card can achieve, hence “floor” as the bottom of the “room” that the card represents. In contrast, point ceiling refers to the maximum value a card can get, like the ceiling of this metaphorical room. These two are connected to the niche-versatile spectrum, in that niche cards often have a high ceiling and a low floor depending on the matchup, and versatile cards’ ceiling and floor are usually closer together in any situation.

One could consider a high floor to be a floor that is higher than many similarly costed cards, and a low floor to be lower than others of its provision cost. On the other side of the spectrum, a high ceiling would be far above its expected value or the ceilings of similarly-costed cards, and a low ceiling would be considered the opposite. If we think about the “room” again, the measure of the ceiling and floor would change the room’s size (the potential values of the card).

A deckbuilder might want to include lots of cards with high point ceilings, but these often come at a cost of high variability and, in turn, low consistency. But too many cards with low point floors lead to a lower point output than many decks that take risks in consistency to get more point value. With most cards, synergistic effects lead to higher ceilings if a card is played with a certain leader or other cards. This is part of Gwent’s card design, and this aspect of the game will undoubtedly improve with expansions.

Even a utility-type card such as Albrich, which technically has a ceiling of 5 points, will usually achieve more value when played in decks that have a few very valuable combo pieces. Although this can achieve value in almost any type of deck by giving access to a key card that was not drawn, it achieves most value when the card it is moving gains the user a disproportionate number of points, and only certain decks contain these ultra-value cards or combos.

When we evaluate cards using this factor pair, we need to look at the point ceiling in the card’s best matchup and the floor in the worst matchup, and think about whether the possible benefits are worth the possible risks - if a card achieves its high ceiling in more matchups than its low floor, it might be worth running. Provisions must be included in the floor and ceiling comparison because they give perspective - if a card has a ceiling of 9 points but is 5 provisions (max 1.8 PPP), it is very different than a card with the same ceiling but 12 provisions (max 0.75 PPP), for example. It is important to both consider the raw value the card is getting as well as the cost required to get that value, since every deck has a limited number of provisions.

One card currently in the game that illustrates a high ceiling and a high floor is Gimpy Gerwin. Technically its floor is 3 (no targets), but practically, it is damaging one unit by 3, in which case it achieves 6 points, a little under its projected value, as well as possible removal of an opponent unit (extra value). However, against certain decks that spawn tokens or other such things (Blue Stripes Scout + Blue Stripes Commando, Arachas Queen, Nekker, even Slave Infantry) it can be worth a lot more, all for a neat 8 provisions.

This card was tweaked in the recent balance patch, as it only needed to hit 2 targets to achieve more than expected value; now, in that case, it will achieve slightly under that value. But if it hits over 3-4 targets, it can be out of control for its relatively low provision cost. This is a problem because the user seldom gets the consequence of a potentially devastating low floor to match its high ceiling - the upsides of a niche card without the downsides.

In fact, the bronze core of many archetypes is based around cards that achieve their point ceiling when played near or with certain types of units (if you control the highest unit, if you have an Elf in your hand, if a card is already boosted, etc.), and those decks contain cards to fulfill those conditions. Synergy-focused decks such as Elves exploit many such cards to maintain a higher-than-average point output.

As was discussed earlier, some of these pseudo-niche cards, particularly those in the Reveal archetype, saw their power level toned down because when played in their archetype, they achieved too much value per provision, and moreover, their point ceiling was being achieved more often than needed, so the “risk/reward” part of it became mostly small risk and high reward. These cards were balanced to either increase the risk by way of higher provision cost, or to decrease the reward.

This concludes Part I of this guide. In Part II, we explore one more spectrum, conditional vs unconditional points, tie that in with the others, and finally, zoom out and look at the big picture to cement our understanding of the many facets of card evaluation.