In the most recent part of the Gwent Pro Tutorial devoted to deckbuilding: ‘Into The Rule of 16‘ and ‘The Rule Of 16 vs Card Drawing Probability‘, a split of the deck into the Preparation and Execution Teams was proposed (1) and the optimal size of teams was roughly estimated (2).
In this article we would like to gather more general knowledge about ‘engine’ and ‘pointslam’ cards which probably would be used to compose both teams. As there is little external material available on the internet (most guides are dedicated to temporary meta), we would pursue our own analysis from scratch (let me know if there are actually some hard-to-google resources accessible). What value does those cards get with respect to round length and what is their best utility? What should be optimal balance between them?
Engines are cards gaining value with time after being deployed on the board. Sidenote: it is funny how the same ‘engine’ term exists in various card games, but means something else: in MTG a card on which combo is built, in Pokémon a thinning/scrolling tool, in Yu-Gi-Oh a synergistic card packet.
We could distinguish some main engine classes with respect to underlying mechanics (Gwent 10.9):
- End-of-turn – cards gaining instant value at the end of the turn – e.g. Viper Witcher Adept, Saul de Navarette, Knight Errant, Gezras…
- ‘One use per turn’ order – cards which bring value every turn when order is clicked, e.g. Griffin Witcher, Pavko Gale, armored Dimun Light Longship, Knut the Callous, Kaedweni Revenant…
- ‘Play a trigger’ – engine which gets value whenever certain condition is satisfied by a played card, e.g. Mage Torturer, Dryad Fledgling, An Craite Blacksmith
- Enemy triggered – engine gets value from cards played by opponent; e.g. An Craite Longship, Arnaghad, Artis…
- Cooldown – cards with cooldown greater than ‘One use per turn’; e.g. Shani, War Elephant…
The groups are not exclusive. For example Reinforced Ballista is both in ‘One use per turn’ and ‘Play a trigger’ classes. Also the list doesn’t include every engine-type card in the game. In this tutorial for the sake of simplicity we would use mostly EoT class for various demonstrations.
End of Turn vs One Use Per Turn
Let’s compare the power of 4p body +1 per turn engines from both classes.
EoT: 6p tempo, 14p ceiling
OUPT (Zeal): 5p tempo, 14p ceiling
OUPT (no Zeal): 4p tempo, 13p ceiling
That’s the general order of power of engines. What should especially be noted is the difference in initial tempo, which may be very important for point gap and reach issues.
The only benefit OUPT could offer over EoT is directed use and synergy with other cards.
Pointslam has two main meanings in common Gwent community use. The first one is the ability to generate proactive points on own side of the board. The second is the ability to do it instantly, with little-to-zero engine impact.
I use the second meaning myself and that’s what we would analyze in this article. Pointslam cards then are ones bringing instant, proactive points value.
Pointslam in Gwent 10.9
After thorough exploration of cards available in the deckbuilder at each provision cost, we could see that there is little to none pure Pointslam cards, especially amongst golds, playing on or above power vs provision curve. Most often a side effect is needed to devlop real value, or the card itself is conditional (let’s take Braathens, Artaud, Tyrgvii, Fucusya).
The pure pointslam peak power vs provision curve rockets from 7 at 4 provision to 11 (Slave Driver into Nauzicaa Sergeant) at 5 provision, and then barely moves up all the way to 11 provisions, with some exceptions amongst Monsters Deathwish cards relying on Overwhelming Hunger (Dettlaff:Higher Vampire as 18 for 10, Brewess: Ritual as 21 for 11). For example Radeyah – a payoff card for singleton deckbuilding – offers only 13 pure pointslam points for 11 provision. Compared with Slave Driver combo, 6 extra provision cost gives 2 points more (inluding possible Assimilate synergies, Radeyah could easily play for less points than Slave Driver).
In this region also two archetype payoff cards from Scoia’tael could reach pointslam value above the curve: Harald Gord and Lake Guardian: Dusk Aspect. A bit similar card in Monsters is Ozzrel at 8 prov. cost, requiring a graveyard setup.
Those rare payoff cards which have to be played late, do not change the general picture – the peak pure pointslam curve is extremely flat in 5-11 provision range. Then in 12+ region there is Simlas into 4x/5x Waylay or other Special card, which would elevate his effective provision cost somewhere near ~16. And surprise: another 2 Monsters cards – Mammuna (instant 20 for 12 on Griffin) and Regis:Reborn (20+ for 13).
And finally, at 14 provision cost there is a neutral card – Renfri, worth roughly 22 points when played as instant pointslam value after Runemage, and even more with passive blessing like Kindness spread over multiple rounds.
As we can see, Renfri is the only high-end neutral card which helps to build the pointslam core as defined in ‘Defending The Bleed‘ article and plays above the Power vs Provision curve.
Another neutral card possibly supporting pointslam core, but at the price of deckbuilding restrictions is Golden Nekker at 9 provision cost. This card could potentially play for around 24 points when used as pointslam, which is significantly above the curve.
Finally, Aerondight, while reactive, could also get considerable pointslam value, especially when the deck is played from blue coin. This special is in a sense R2 push payoff card, analogous to Lake Guardian, or Regis:Reborn as archetype payoffs. As we see, this type of cards is one of few ways to build the core.
Engine Value Over Time vs Pointslam
- +1 Per Turn
Let’s take a +1 EoT engine and compare it with characteristic pointslam milestones. The initial point value would be varied.
Out of 5 plotted lines, probably the most interesting is the one starting at 4 base points, which resembles typical cheap bronze engine.
End of turn engine by definition plays at least as 1 turn value. In the case of play and pass in the next turn, we are at 2 turns value already. EoT+1 engine tempo is then only worse by 1 point than best pointslam 4 provision bronzes (Nekker Warrior) and breaks even when. To get the value of Nauzicaa Sergeant, 6 turns are needed. In a full 10 card R3, the engine could outpoint Radeyah by 1.
- +2 per turn
A +2 per turn engine in a long round equals the value of strongest pointslam cards when played first from 2 base power. Higher base power +2 engines obviously outvalue even strongest pointslam cards.
Windhalm of Attre is a model example of +2 per turn engine which could be played early. After 4 turns the card returns provision, playing for the same points as worth 8 provision total Renfri Gang pair.
The relative value of Windhalm getting answered is worth no less than 18 points; a tool stopping or countering Windhalm would be worth as much as best pointslam cards.
Even cheap Witch Apprentice if for example played right after Mammuna setup in the first move, could reach Regis&Renfri value.
- +3 per turn
The model example of +3 per turn engine is Saul de Navarette, but many powerful engines could be assessed by using similar curve, even if working differently. For example Roland Bleinheim or Townsfolk are engines which could be visualized as +3 per turn. Sometimes they play for less, sometimes for more (also obviously don’t buff at the end of turn), but just powerful engine picture is enough for many assessments.
Saul plays as 10 tempo and evens Radeyah value after just one more turn. After 6 turns, Saul is on par with Renfri&Regis. The model Saul has 34 points ceiling (in practice some coins would be spent, at least in the last play).
- Playing pure pointslam in Gwent 10.9 environment is unconceivable: +1 per turn engines are reaching value of mid pointslam golds in a long round, +2 evens best high end cards and +3 outvalues them by at least 10 points. Decks close to perfect pointslam (Renfri Imprisonment), have to run many control tools. That’s the reason behind leader choice, Alba Armored Cavalry and finally Geralt: Igni. While these tools may look like a waste of points in pointslam mirrors, against engines those are absolutely necessary; for example Igni value on model Saul would be equal to 36 points.
- The most powerful engines are hidden meta definers – it might seem weird after Aerondight + Nekker and Renfri metas, but its true. Strong 3+ points per turn engines potentially bring most points on the board and the outcome of any game would come down to them getting answered or not. If engines are answerable, then cards countering them efficiently become meta holy grail. If greedy engines are simply too weak and vulnerable, then best pointslam cards start to emerge as meta staples. Examples of engine definers would be Self-eaters and Bloody Mistress in Relicts meta, Raffard’s Vengeance in Alumni meta, Sigvald+Knut in Self-wound or Mysteries of Loc Feiann, Damsel in Distress and Roland Bleinheim in the current Gwent 10.9 meta.
- Comparing engines with pointslam counterparts around the same provision cost, we could see that only 3 or 4 turns are usually enough to even out the field between pointslam and engine of same provision cost; Scenarios develop then value comparable with Renfri or Regis, Dryad Fledgling value equal to Nekker Warrior, Antherion value equal to Gan Ceann. The main downside of engines is the possibility of giving value to opponent’s control cards, especially the ones from the Execution Team. Then usually the stronger the engine effect, the bigger point loss with respect to the pointslam counterpart. Putting low floor engines in the execution team is then obviously very risky.
- Denial/removal of engines is a hidden value and removal value scales in the exact same way as engines when opposing them. What matter then for meta shape and the choice of played cards is the trade between engines and control; does shooting down an engine generate positive points, or rather break even or down in points/provision?
- At the first sight it might seem like all engine R1 hand is the best preparation team to fight for Round 1. But in reality, not all R1 hand is playable; on average 50% of execution team is drawn in R2 and it would be optimal to not mulligan out at least the strongest pieces. Then let’s say 4 cards have to be saved. You could contribute no more than 7 turns by yourself to an engine value. Then a set of 1/5, 2/4 or 3/3 pointslam/engines(or control) would be optimal to maximize points. The deeper R1 could go, the better it is to include more engines/removal with respect to pointslam to fight for R1. At the same time the shorter R2 and R3 get, the better poinstlam cards become.
- The natural composition of long R3 execution team would be 3/7 or 4/6. The exception would be very aggressive decks, which may need more pointslam for better late R2 and short R3 value. Again, the main criterion is card quality, and split should not be rigidly obeyed.
- When playing against removal in a long round, it is advisable to place your weak engines before strong ones. Imagine that you run one +3 engine, one +2 engine and one +1, and your opponent has only 2 answers. Then if placing weak engine first, opponent has to wait for stronger ones to appear. The suspension could last many turns, and refraining from locking +1 engine is hard psychologically. Only later on +2 and +3 engines get deployed; at this point maybe opponent already lost nerves and locked +1. No need to hurry. On the other hand, when the strongest engines are deployed first, they both get answered and the round for +1 engine simply becomes shorter. Even if opponent do not take the bait, placing weak engine first is worth 2 more points.
- Shortening R3 by a single card denies one turn from all non-answerable opponent’s engines, as well as non-answerable own engines; if opponent’s engines would make +8 per turn in long R3 after full development and your own +2 per turn, then each card shortened from R3 brings you +6 carryover.
Thanks for reading! The graphs and remarks presented here are only to invite you to own discoveries and thoughts. The topic of various card classes and sequencing would be continued…