Thursday, April 18, 2019

When to scream for nerfs

The title is a bit dramatic :), and I think we can all agree that nerfs (and buffs) are important. But, when should something be nerfed? Are the commonly-heard arguments for nerfing good ones, and is there a set of guidelines that suggest when something should be nerfed?

I had an interesting conversation with Travis Johnson on Facebook a while back about Ion Y-Wings, specifically 5 Y-Wings with Ion Cannon Turret where 4 have Veteran Turret Gunner. I argued one list has to be the best list at jousting (assuming for now that list is the best jousting list, which seems to be dubious), and it doesn't make sense to always call for nerfs for the best jousting list. Travis argued number of attacks and the ion effects feel bad to play against. Someone (I can't remember if it was Travis or someone else) also said the Y-Wings are too easy and boring to play for how good they are and thus should be nerfed.

These are things I've been thinking about for a while. What are the easy lists or hard lists? Should easy lists be nerfed? Also, what is this "burden of execution" thing that gets thrown around and how does that relate to nerfs? What are other good reasons to nerf a list?

My conclusion is whether a list is "easy" or "hard" isn't a good argument for whether it should be nerfed. Furthermore, lists whose success depends more on the opponent's decisions than the player's tend to be weak for tournament play.

The ships and lists that deserve nerfs are the ones that lack counterplay, where their success depends mostly on their player and less on their opponent's decisions. Signs that a ship lacks counterplay include out-jousting most less-maneuverable ships and changing their position with perfect information. These aren't hard-and-fast rules, but are rather warning labels that suggest a closer look.

Aside: Sometimes people consider calls for nerfs as attacks on the player's skill. Lists that lack counterplay scale really well with player skill because they depend more on their player's decisions than the opponent's. Winning with a list that lacks counterplay may mean the player is high-skilled, not low-skilled. The list may still need to be nerfed for the health of the game.

Finally, some game effects are balanced or weak but aren't fun to play against. I think there are reasonable arguments to remove these from the game by nerfing them, but the case to nerf these is harder to make. Different people find different things fun, and they won't be too common if they're already balanced or weak.

Easy vs. hard lists

First, let's consider lists that are easy to play or hard to play. Lists that are easy to play have a natural advantage over a long tournament, and it feels good to reward players that master a hard-to-play list.

But what are the hard lists? I often see the idea that fragile arc-dodgers are hard to play while stodgy jousters just 1-straight to victory. A strong jousting list does well against beginners and arc-dodgers are much weaker in the hands of a beginner. But for this standard, we should consider players who make the cut or are on the bubble of major tournaments. After all, the mental tax argument isn't relevant for casual night games.

Note: I usually use the terms arc-dodger and jouster as roles ships play within a matchup and not a universal feature of ships, but here I'm going to refer to less maneuverable ships as jousters and more maneuverable ships with great repositioning options as arc-dodgers.

And at the tournament level, jousters and arc-dodgers are usually equally hard to play. If anything, jousters are harder to play.

Sure, the arc-dodger has more options to consider. But to catch the arc-dodger, the jouster also has to consider the arc-dodger's options with probably less experience flying those ships! Who has an easier time figuring out where Advanced Sensors Guri can go: the player who's practiced Advanced Sensors Guri and is looking actively at the dial considering her options, or the player flying X-Wings playing against it for the first or second time?

Sure, an arc-dodger might suffer crippling damage after one careless maneuver. But it's the same for jousters. They might not die right away, but one bad maneuver is all it takes to lose the game. A bad maneuver might do anything from taking multiple ships out of the fight to simply having some ships just out of range in the initial joust and thus lose the damage race.

For the Toronto System Open, I flew Wedge, Luke, ARC Norra with Leia, and a Z-95 with Selfless. Three very efficient jousters and a Selfless chaff that hopefully gives its life for the others. I went 3-3, and although all of my losses were close, none of my games felt easy. My first game against 5-ship Rebels started with me initiative-killing two ships in the first two rounds in exchange for Luke. One round of bad maneuvers, bad target priority, and bad dice luck later, I lost Wedge and the Z-95 that round and by 1 MOV when time was called a few rounds later. Similarly, my third game was a loss against Resistance 5's. One careless maneuver put Luke out of the fight for two crucial turns and put me in a position need decent dice luck on the last round of shooting. Some of my wins followed a similar pattern: a good opening would turn into a much closer game after I made a mistake.

If anything, arc-dodgers are more forgiving to fly. Losing an arc-dodger means you can't initiative-kill weakened ships as easily, but at least the rest of your arc-dodgers will have more room to play. For jousters, each firing arc they lose means a smaller chance of catching their opponent in the future, and it's much harder to outplay for free shots with jousters to recover from a bad decision or bad luck.

In the last few rounds of that first game at Toronto, I had to out-fly a generic T-65 and a Sheathipede with an ARC to win. That was too much for me to handle. On the other side, after seeing PhilGC do crazy things with his two-ship lists, I tried out his Fenn Guri list. I have nowhere close to his experience with two-ship lists (never flown in 2.0) or Guri (only flown a few times in 2.0, although I've flown generic StarVipers a few more times). In my first game, Fenn was crippled in the first round of shooting after one of my patented terrible barrel rolls took him out of range 1 but not out of arc. In the second, Fenn died in the first round of shooting to two 2-dice attacks at range 2 and 3. But I still had Guri. She danced around, forcing and capitalizing on small mistakes, avoiding shots while hitting back. A last-round misplay by my opponent put me 1 MOV ahead in the first game. In the second, Guri bumped to deny shots and initiative-killed a crucial ship to open things up for a full wipe.

If jousters or arc-dodgers don't determine a hard or easy list, what are the truly hard lists to fly? Those are the lists full of low-maneuverability ships that consistently lose the joust, or fragile high-maneuverability ships that can't reliably arc-dodge or put out enough damage before they're inevitably caught. In other words, these are lists that have tons of bad matchups. The hard lists are the lists that are weak. And similarly, the easy lists are the strong lists which have good matchups across the board.

And if that's the case, "hard" and "easy" don't give us any information on whether a list should be nerfed. Yes, list that has good matchups against most of the field is too strong/easy, and it should be nerfed. Yes, a list that has bad matchups against most of the field is too weak/hard, and it should be buffed. If they mean the same thing, then a list being "hard" or "easy" isn't an argument for it being strong or weak. We'd be saying a list should be nerfed because it should be nerfed.

Burden of Execution

Another popular term is "burden of execution", where the common argument goes that lists which place the "burden of execution" on their opponent should be nerfed. I think it may be used in different ways, but for this article, I'll say this means a list's success depends more on the opponent's choices than your choices.

As my attention turned to the Hyperspace Trial season after the Toronto System Open, I tried Wedge, Luke, a generic U-Wing with Leia, and a generic X-Wing. Again, I had two efficient jousters, the hyper-efficient Leia, and a strong filler ship. I got some wins with it, but my friend Marc (top 4 Toronto System Open, master of memes, and the noble sacrifice that saved us all from the Toronto Tallon Roll) was consistently beating me with his 4-ship TIE Salad.

After banging my head against the wall to try to figure out this list, I finally realized what the problem was. My best overall and end-game ship is Luke (second is Wedge). My most fragile ship in proportion to the damage they deal is Luke (and Wedge). That meant Luke (and Wedge) is the priority target. And unfortunately, I couldn't find a reliable way to keep them safe.

In other words, that list put the "burden of execution" on my opponent. If they correctly chose to kill Luke first, I was in trouble. Otherwise (some people got distracted by the U-Wing), I'm favored to win.

The problem is, I have no control over my opponent's target priority. A good player like Marc will make the right decision every time. That made the list highly unreliable and a bad choice for tournament play. Rather than a reason for nerfs, lists that place the "burden of execution" on their opponent tend to be weak.

Counterplay

If lists that win or lose mostly by your opponent's decisions are weak for tournament play, then lists that win or lose mostly by your own decisions are too strong. And it's precisely these lists that are most concerning for balance. If the success of a list is only determined by its player's decisions and not by its opponent's, then it'll be unbeatable in the hands of a great player. X-Wing is a two-player game, and we usually play these games to interact with the other players. Both players' decisions should contribute to the end result, or in other words, both lists should allow for counterplay (I prefer "counterplay" to "agency" for this concept).

Of course, lists don't have to be literally unbeatable to deserve nerfs. If the decisions your opponent has to make to win are too hard to reliably execute (perhaps a different use of the phrase "placing the burden of execution on the opponent"), then the list doesn't have enough counterplay. Maybe he has to put his ships in multiple exact spots for multiple turns, correctly guess the exact maneuver you picked, or outfly you so much that he gets a full round of free shots even when his ships aren't that maneuverable.

A list can lack counterplay for two interconnected reasons. First, a list can have an overwhelming numerical advantage. This might be a list which simply deals or avoids too much damage. As an example, the generic TIE Phantoms (with Jan 2019 points) are among the most efficient ships in the game even without Juke.

Second, a list can have an overwhelming maneuverability advantage. These are ships with high initiative that can change their maneuver or reposition before executing their maneuver after seeing where the opposing ships are. Normal arc-dodgers have counterplay in that they are committed to the maneuver they dialed. Opponents can turn off their ability to reposition by blocking them. Ships that can reposition or change their maneuver before executing their maneuver are almost impossible to block, so there's no universal mechanic that any list can use to pin them down.

These are interconnected problems. A list full of shuttles may have a numerical advantage if their dials aren't considered, but they're easy to get free shots on. TIE Phantoms may have worse stats compared to shuttles, but their numerical advantage is partially based on their above-average maneuverability. In fact, they joust better than most less-maneuverable ships, a classic sign of a ship that lacks counterplay. Similarly, the Grand Inquisitor can take Supernatural Reflexes just like Kylo Ren can, but Kylo's stats make him more effective at dealing enough damage to get ahead on points.

It's worth mentioning hard-counters here. Hard-counters are effects which are much stronger against some lists. Even though they are weak against certain lists, lists they hard-counter may not have enough counterplay against them.

Similarly, nerfs can also be justified if the gap between top-tier and "normal" lists is too large. Top-tier lists may be competitive against each other, but may lack counterplay against most lists. This was the situation for most first edition, and most players could agree this wasn't an ideal situation. There will always be good lists and bad lists, but the bad lists should still stand a chance against good lists with strong play.

Balanced but feels bad

Finally, there are effects which may not win a lot, but aren't fun to play against. The most common examples are effects which take away choices players can normally make, like ion effects or multiple stress effects. Effects that reroll your opponent's dice can also feel bad because of loss aversion.

There's an argument that these should be nerfed. X-Wing is a game, and games are played for fun. If you're not making choices in a game, you might as well not be playing. If these effects reduce the amount of fun in the game, then a nerf that reduces their prevalence can make the game more enjoyable.

That said, there's a couple reasons why this is a harder argument to make. First, different people have different ideas of fun. Some players may enjoy having a control element in their list, and some people may not mind playing against them when balanced. Second, if these effects are already balanced or weak, they're not likely to become popular. They might not be widespread enough to significantly impact the fun of the game.

Keep Calm and Scream for Nerfs

With all of this in mind, we now know some warning signs for when we should scream for nerfs. We should be on the look-out for ships that out-joust most of the less-maneuverable ships. We should also keep a close eye on ships that can change their positioning after having full information of where most enemy ships moved. Finally, we should also keep an eye out for effects which take choices away from players.

I also hope we can give more credit to players who prefer swarms or jousting lists. Rather than being easy no-skill lists, jousting lists require good decision-making to win with and are probably harder and less forgiving to fly than aces. A recent meta analysis found there are slightly more I4, I5, and I6 ships being played than lower-initiative ships. That's the opposite of what's expected since lower-initiative ships tend to be cheaper so more fits in a list. I hope FFG can show more love for these lower-initiative ships (maybe by raising the price of higher-initiative ships) so jousters and swarms will see more play.

Tuesday, April 2, 2019

The Value of Heroism

is 2 points.



Heroic is a fun card for theorycrafters. It's cheap, can have a big effect, but rarely triggers. Its benefit is pretty easy to quantify. Many have done the math on this upgrade, with Zombie Squadron's article as one of the more recent examples. Still, opinions seem to be mixed on whether this card is good. Some people say the benefit is strong enough, some say it triggers too rarely to matter, and some say it's only good because it provides insurance against bad outcomes.

I want to answer two questions to fill in some gaps and settle the debate. First, exactly how much does Heroic improve a ship? Second, how many points is Heroic worth? To answer these questions, I calculate how much more effective a ship with Heroic becomes. This is the best metric to see whether Heroic is strong or weak.

Most of the numbers were generated from the X-Wing Probability Calculator. I'll average the benefits using these assumptions:


Attacks received are always assumed to have a single mod, and I also assume 2 attacks received per round for focus on defense.

Technical note: the damage dealt scaling turns average damage dealt against a defense profile into "average attack equivalents". These are calculated using the meta assumptions on attack profiles. For example, 1 damage against a 0-agility target is only half of an "average attack equivalent" while 1 damage against a 3-agility target is worth one "average attack equivalent". Damage is scaled so a percent increase in damage dealt to a 3-agility target is worth the same as a percent increase in damage dealt to a 0-agility target (before adjusting for the frequency of these defense types). Without this adjustment, percent increases in damage dealt to a 3-agility would be worth less because it starts from a smaller value.

Heroic Offense

The graph below has the percent increase in damage dealt that Heroic adds. The average benefits at the bottom average across the meta assumptions including range bonuses, while the numbers above the line are just for those raw dice.



As expected, Heroic adds the most benefit for 2-dice attacks. 2-dice attacks improve by 6%, while 3-dice attacks are only improved by 1.5% and 4-dice attacks gain a negligible benefit. This means the average benefit including range bonuses is roughly 4% for a 2-attack ship like an A-Wing, while the offensive benefit for X-Wings is only 1%.

The percent benefit of Heroic on offense is about the same with or without focus. Of course, the raw damage increase is larger with focus because it starts from a higher base. Intuitively, this makes sense. Heroic gives you a retry, and the strength of that retry is just the strength of the base attack. The benefit of Heroic as a percentage of the base attack will just be the percent chance it triggers, and focus doesn't affect that at all.

This is not part of the analysis, but keep in mind Heroic doesn't help at all if you have already have a Lock on the target except by improving the chances to keep the Lock for the next turn. If you're frequently taking Locks, then Heroic's value on offense goes down.

Heroic Defense

The graph below has the percent increase in durability that Heroic adds. The average benefits to the right average across the meta assumptions including range bonuses, while the numbers left of the line are just for those raw dice.


Again, Heroic adds the most benefit for 2-die defense rolls and 2-agility ships. If we assume a 50% chance of having a focus in a round for defense (as opposed to saving it for offense or not having a focus token at the start of the round), a 2-agility ship like the X-Wing sees an 8% increase in durability while a 3-agility ship like the A-Wing sees a 5.5% increase in durability.

Unlike its benefit on offense, the benefit of Heroic on defense improves when the ship has a focus token. Intuitively, this also makes sense. For the easiest case to imagine, consider a 1-hit attack. When the defender doesn't have a focus token, this will only deal damage when the defender rolls only blanks or some combination of blanks and eyeballs. Heroic reduces the chances of rolling blanks, but not the chance of rolling some combination of blanks and eyeballs. When the defender has a focus token, this will only deal damage when the defender rolls only blanks, and Heroic reduces all of that chance.

Similarly, Heroic is also stronger against weak attacks than strong attacks for a similar reason. A weak attack relies on the defender blanking out to push through meaningful damage. A strong attack can still push damage through even if some evades are showing.

The Value of Heroism

We can use the simple method in my last article to calculate the total benefit of Heroic. We add the offense benefit to the defense benefit and divide by 2 to adjust for focus fire. It turns out the benefit for A-Wings and X-Wings are pretty close; Heroic makes A-Wings 4.8% more effective and X-Wings 4.5% more effective.

We can multiply these benefits by the ship's cost to find out how much Heroic should cost to keep the ship's efficiency the same. Here are some examples based on the January 2019 point costs:

  • Blue Squadron Escort A-Wing: 1.5 points
  • Tallissan Lintra: 1.7 points
  • L'ulo L'ampar: 1.7-1.8 points (based on how often he's stressed)
  • Blue Squadron Recruit X-Wing: 2.1 points
  • Nien Nunb: 2.5 points
  • Poe Dameron: 3.1 points

As we can see, Heroic should cost about 2 points on average to keep the ship's efficiency the same. If Heroic is added to a ship with other upgrades, the cost of the other upgrades should also be included and further increases the value of Heroic.

It only makes a small difference, but the true value of Heroic is based on the true value of the ship and not the official FFG point cost. Most of these ships are costed pretty accurately, but Tallissan and L'ulo are some of the most underpriced ships in the game. If they were priced more accurately (and with more precise modeling), Heroic adds about 1.9 points of value for Tallissan and 2.4 points of value for L'ulo. Poe might also be a bit overcosted, but the difference is small with R4 Astromech and he already benefits greatly from Heroic.

Overall, Heroic would be fair for most ships at 2 points and it's a steal at 1 point. The efficiency of Heroic is magnified because it's a cheap and unlimited upgrade that most of the popular Resistance ships can take. You can bring 4-6 copies of Heroic in a list and get 4-6 extra points of value.

That said, this might be a good thing. The point of faction identity talents is to make each faction feel different, and they can't do that if no one uses them. Heroic differentiates Resistance from the other factions and doesn't hard-counter specific archetypes (like Treacherous), greatly limit list-building (like Ruthless), or restrict the playstyle of lists that take it (like Selfless, or Dedicated to some degree). Fearless and Fanatical are other examples of what I think are well-designed faction identity talents in that I wouldn't mind if every Scum or First Order list included those (I wonder if Dedicated is interesting enough to qualify or if it would just get old fast). I'd be happy to see these talents undercosted so they see the table more frequently and do their job of making each faction play a bit differently.