Month: October 2017

[OTC] First Builder Advantage

I’ve featured Offworld Trading Company analysis on the site before, as its relatively simple rules and deep emergent depth make the game into an effective lens for highlighting cross-game strategy fundamentals. This post is the first of three covering topics I’ve pulled from philothanic’s coverage of a recent 4p FFA on Ceres. First and today, we’ll see an example of an early head start in development and tempo chaining (temporarily) into a maintained lead. Later in the week, we’ll look at how three players make the seemingly-identical choices to construct an Offworld Market yields unequal value due to differences in their positions. The third post in the series will let us look at distributed value being less impactful than concentrated value, as one player buys up a little bit of each other player’s stock instead of focusing on a target.

So, on to our first topic, the chaining of tempo/development advantages to maintain an early lead (at least for a while). We’ve seen this theme before in prior posts on Overwatch and Stellaris respectively.

Unfortunately this post may seem like we’re picking a bit on poor Adorfield, as he was for various reasons at different times a bit too slow to upgrade when he could have and was punished. Let’s take a moment to acknowledge that he’s up against some of the best players of the OTC community, players quick to close their windows of vulnerability and faster to even inadvertently take advantage of others’ missteps.

Near the start of the game it is critical to upgrade to HQ2 quickly to gain extra tile claims to turn into resource production and profits. Adorfield got too tricky for his own good, at HQ1 buying up an early food stockpile in the knowledge that its price would inevitably rise. We’ve actually had a previous post where the same player used a similar tactic that did not pan out. Not a bad idea in and of itself, but here he could have sold some of the food in exchange for a faster upgrade and more tiles to leverage for profits:

Later on, Adorfield and Portia at HQ2 are BOTH punished for failing to upgrade to HQ3 when they had a window to do so. DeathTacticus, already a full HQ level ahead of them, beats both to it and ticks up to HQ4. Since he buys some of his upgrade resources off the market, their prices rise, increasing the costs of Adorfield’s and Portia’s upgrades. By being first to (yet again) click up, DT is able to maintain his early advantage and put the onus on the others to overcome inflated prices on the resources the other players need to follow suit. Sometimes, the inflation of market prices can punish the early upgrader, if another player is producing excess of those resources and can then sell off the surplus for extra cash, but in the early stages this is uncommon, as all players are fighting to upgrade quickly rather than overproducing construction resources for profit:

Adorfield has a lot of money on hand, and uses it to buy up the first half of his stock and then upgrade. This is safe but not necessarily optimal. He could upgrade himself and invest in buildings, patents, and production optimizations which he could cancel out of in case he needed that cash to defend his stock from a majority buy. Or he could sit at HQ4 and try to find a window to use his cash stockpile to tactically strike another player with a buyout. Also of note, Adorfield had a chance to take the upgrade earlier but avoided taking it, presumably to not be a target as the only HQ5 player, given that he was going unnoticed so far. But the guaranteed concrete advantages of upgrading could have been well worth the non-guaranteed risk of being dogpiled as a threat:

[All] Don’t position against a wall (without good reason)!

Today we return to the site’s roots, looking at a general strategic idea that exists across the gaming world: “Don’t position against a wall!” I’ve most recently remembered this idea from reviewing and playing Overwatch games, but it has cropped up many times before. Sometimes the wall is physical, sometimes it is metaphorical, but by definition it limits your available moves. This is a gift to the opponent! It makes their job easier. We want to do the opposite. As I’ve put it before, we want to “put the onus on the opponent.*” We want to make their job harder, to give them opportunities to misplay or guess wrong. Let’s look at some examples.

(11:54-12:18) First I’ll take a look at an Overwatch clip, though the ideas extend very naturally to games involving realtime movement prediction. It’s here that the concept of not positioning flush against a wall most recently returned to my thoughts. Here is a brief example that cropped up while I was reviewing and analyzing a set of replays from a player’s run to diamond rank. The Ana, by positioning right up against a wall, takes hits from all 3 of Genji’s shuriken. If she had a little more leeway to play with, she could force Genji to divine her movement, putting the onus on him to guess correctly in order to land his damage:

The example above has analogues in other FPS and also arena games, like League of Legends. For example, a Thresh chasing down a champion towards a wall will wait to hook until the enemy is up against the wall, reducing the directions they have for juking But you might be surprised to realize this idea comes up even outside of real time movement-prediction games. In chess they often say “A knight on the rim is dim,” referring to the weakness of the knight when on the edge of the board. A centralized knight could move to 8 squares, but on the edge will have only two options. It’s so bad that the rim knight can be completely “dominated” by a bishop, as presented by YouTube commentator MatoJelic in this clip:

(19:58-20:26) In the Overwatch example the wall was a physical wall. In the chess one, the wall is almost physical, consisting of the edge of the board. Now let’s look at a Magic: the Gathering example. In this clip from the recent Week 7 of Vintage Super League season 7, Rachel Agnes highlights a good example of a metaphorical wall restricting a player’s options. Rodrigo Togores has only 1 life left, making him unable to pay the alternate cost of Force of Will (card image below clip) without losing the game. Having even one more life would require his opponent to play around the possibility of a no-mana-cost counterspell, but as-is he’s against a wall. The opponent knows that if he does have the Force, he must pay the full 5 mana for it or not use it at all. Rachel goes on to share the amusing tale of a past an opponent of hers who threw away a winning game by needlessly paying their last point of life to counter a spell unnecessarily:


* Aside: In the past, I’ve used the phrase “put the onus on the opponent” to talk about minimizing risk while comfortably maintaining a gamestate where the opponent is disadvantaged, where if nothing changes the status quo will bring you more advantage or a win, forcing the opponent to take on risk in trying to “make something happen” that might upset this trajectory.

[Civ V] Pop micro on growth turns for extra hammers

Civilization V NQMod streamer Yoruus explains how he manages to squeeze an extra 2 production hammers out of his city on a population growth turn. In Civ V, city yields are processed in an order that starts with food, which means a new citizen can be born and start working a non-food tile to contribute to income on the same turn it is created. Further, with the NQ mod unassigned citizens (“laborers”) provide two hammers of production rather than the unmodded game’s one. By setting his city to “focus production” and leaving an unoccupied hill (which can be worked for 2 hammers), Yoruus ensures that the new citizen will go to the hill and provide an additional 2 hammers this turn towards the production of the Temple of Artemis wonder he is racing to build before any other player completes it.

Contested gamestates are won at the margins, a theme that crops up just about everywhere. Eeking out a mere 2 extra hammers may look like nothing, but it can push one over a threshold of effect by changing a 3 turn build to 2 turns. And even more subtly, it can alter the threshold of effect down the line, as the carryover of production overflow will influence the timings of later builds. Such tiny advantages accrue and over the course of a long game can actually move up a city’s entire production schedule by several turns, bit by bit.

[OW] EnVyUs punts a critical moment in Overwatch Contenders Grand Finals

EnVyUs misplayed a critical moment fighting to take Objective B on Temple of Anubis. They ended up winning the map anyways, but the punt give their opponents a window in which to regroup and potentially hold. Times mentioned refer to the video sub-clip below.

The situation: EnVyUs (in blue and on attack) is leading due to victories on prior maps and has just taken Objective A. Eager to wrap up the game, they quickly make a move for the second point. Tracer and Sombra’s mobility puts them well ahead of their team. Working together, the two score a pick on the FaZe’s Sombra (at 14:43 in the clip). Note at this moment that EnVyUs’ Lucio is at 86% of the way towards his Sound Barrier ultimate, a key tool for teamfights. Further, though they don’t know it, their opponents’ SB is at only 52%.

Hungry for blood from the pick, EnVyUs’ Effect (on Tracer) overcommits, quickly dying at 14:48 (aside: Effect overpushed and died alone for no reason a number of times this match, e.g. as Reaper shortly after the end of the portion clipped here). The overcommitment also baits her Winston, who splits from the still-approaching team with leap to die alone on point at 14:51. A mere 3 SECONDS later, their Lucio’s Sound Barrier becomes ready, and he pops it immediately in the 4v5 (though its shield on the far-back Widowmaker is essentially irrelevant, so it’s basically a 3-man SB). The attackers kill Genji at 14:58, then burn D.Va’s Self Destruct at 15:01 for no kills (but the first 1/3 checkpoint of cap time), followed by Sombra’s EMP at 15:07 right before she dies. Despite the hype from the commentators, EnVyUs gains no real traction here and they lose control of the point. The short respawn distance for the defenders means they are able to hold and regroup. The attackers burned many ults while the defenders used only one (a late Self Destruct from their own D.Va at 15:20). FaZe is left with some time to defend in relative comfort with their ult advantage, forcing Envyus has to burn some time before finding another credible opportunity to push.

With better communication and less bloodlust, EnVyUs could have delayed their dive for just a few short seconds and been able to use a 6-man Sound Barrier against the 5 defenders that had no SB of their own available. It almost certainly would have resulted in a clean wipe and a map win. This one mistake cost them time, dragging the game out for an additional 4 minutes. They gifted FaZe a chance to stay in the game. Though EnVyUs did in the end take the map, it’s important to learn from victories as much as from defeats, and the dives from Tracer and Winston before their one-sided Sound Barrier could come up was a key mistake at a critical juncture.

[HoMM3] Quicksand turns the tides

HoMM3 player Chris67132 stylishly navigates a tough early spot with style and Quicksand, a normally underwhelming spell he picked up from a level 2 spell shrine. Chris knows that a good or bad opening causes ripples that influence the whole game.

(10:00-11:00) With his main army positioned in the east with few targets in range, he faces a choice between burning multiple turns without deveoloping or putting everything towards beating “lots” of demons. Wasting time is asking to slowly lose down the line, so puts everything towards beating the Demons, knowing that he’ll also need a bit of luck with the random placements of his Quicksand to stand a chance. He also remembers from his last run that the wandering monster stack there was merely a “pack,” (10-19) one category smaller than “lots” (20-49), which gives him reason to believe the number demons will be closer to 20:

(11:03-15:40) The informed guess about the numbers proves true, with exactly 20 demons appearing in the group, the lowest amount possible for lots. He gets enough luck with quicksand to impede several Demon stacks over the course of the fight. He buys more time for his shooters by tempting demons to attack singleton out-of-the-way gargoyles. All told, the demons are delayed from closing with the shooters long enough for the ranged damage to take out almost all of them, cinching the much-needed win. The victory will allow him to him take the now-little-guarded neutral city early:

[SSBM] Commentary moments on Sixx’s Link at Heir 4

Sixx had a great run at the Heir 4 SSBM tournament with his trademark Link. Link is generally (and rightly) ranked as a B-tier character. In theory, he should be outperformed by faster and more tryhard tourney choices such as Fox and Shiek, but in practice a Link specialist can put on a real show. By leveraging Link’s strengths, Sixx battled his way to 9th place out of 476 entrants in the singles bracket with a string of brilliantly entertaining matches. Strong Link play is uncommon, and many commentators are ill-equipped to provide useful insight into the character. It was a rare pleasure to hear on-point discussion from Lolo and Fuzzyness. I’ve pulled three good moments of analystical commentary from one of Sixx’s early matches.

(00:08-00:16) One common way to try and punish an attacker as the defender is to block a hit with shield and then immediately drop the shield to counterattack while the opponent is still locked in the end of their attack animation, but after their attack hitbox has deactivated. Link’s forward smash is resilient against this tactic because the hitbox remains for a long time. The usual trick of dropping shield quickly doesn’t work against this strike due to its lingering hitbox, which will still be live to contact the opponent after their shield comes down. This high-duration hitbox protects Link despite him being a slow character:

(00:27-00:38) Crouch cancelling (CCing) is a SSBM technique where a player holds down on the movement stick while taking a hit, causing them to be knocked back far less than normal due to their character’s friction against the stage floor. CCing is useful because it allows one to take a hit without being knocked out of range to counterattack. Most characters’ down-air attacks hit opponents downwards, making them ideal attacks to crouch cancel. Link’s d-air is unusual in that it instead delivers a powerful popup. Since crouch cancelling depends on stage friction, Link’s popup d-air cannot be CCed. As the commentators point out, Sixx takes advantage of his unusual d-air to get in a one-sided hit on Peach, as she cannot reduce its knockback with a CC:

(04:22-04-55) In this last clip I’ve taken from the match, we hear how Link’s mobility is a weakness. Perhaps his biggest single weakness is the extreme slowness of his jumpsquat, meaning he has a long animation to go through before lifting off from the stage when he tries to start a jump. He is also slow when coming back to stage from the air, as his jumps and attacks have a lot of landing lag before he can start another move. He is a sitting duck to be punished during these windows of forced inactivity. On the bright side, Link’s attacks apply high damage to shields, which can help limit the opponent’s options and provide some safety…assuming the attacks are placed well enough to strike the other character or their shield in the first place:

[AoE II] Mathcrafting a villager gather rate formula

The following video by YouTuber Spirit Of The Law is one of those few-but-beautiful videos that is tightly targeted and site-appropriate, being chock full of great game analysis all the way through. Spirit, an Age of Empires II analyst, uses math and in-game testing to arrive at data-driven answers to strategy questions. Here he gets to work to create a useful formula for estimating villager gather rate, previously nebulously-understood. After arriving at a formula, he goes a step further, applying it to answer some otherwise-unanswerable questions:

[Stellaris] Is Mastery of Nature excellent or terrible?

Let’s hear from two Stellaris players on their very different evaluations of the Mastery of Nature (MoN) ascension perk in Stellaris.

Macsen likes taking MoN early and often in his games. The perk provides for free all tile-blocker-clearing techs and a discount towards tile clearing. In the early game, when all types of income are lower, the free techs and cheaper energy/mineral unblocking costs make a real difference. The techs alone tend to cost upwards of 20 months apiece in the early game. As for the yields, note that at the moment here where he’s picking MoN, his energy income merely +3 and his mineral income is +51. Most tile blockers cost 100 energy and minerals to remove. This means MON currently reduces the cost by 16 months worth of energy and 1 month worth of minerals. Macsen’s energy stockpile will soon be quickly drained by the costs of colonization and fleet maintenance, and he’s already using up his mineral income month to month, so he will definitely be able to make use of the discounts.

Tokryva disagrees, In a video (partially) titled “Mastery of Nature sucks,” he shares his reasons for strongly disliking the perk.

MoN provides a a 50% discount on the energy/mineral costs of blocker removal, but Tokryva does not value this bonus highly. He’s right that the discount is the weaker half of the perk:

Tokryva population growth is only limited by tile blockers when there are no unblocked and unworked tiles available. Since planets do not fill their unblocked tiles very early, Tokryva does not feel rushed to clear tiles:

Tokryva points out that Mastery of Nature’s free techs can be acquired the normal way, while the other ascension perks provide benefits that do not become obsolete. “Every tile blocker removal you can get by just researching. Why do you need them up front?”

In a similar vein, while there are many tile blocker types, they vary based on planet type, and in the early game the player will only be settling a few types of planets and therefore face only a few types of blockers.

Overall, Tokryva is completely right that Mastery of Nature falls off to the point of being essentially useless later in the game. However, his ascension perk choice came later than it could have because he split his tradition picks between two trees. Amusingly, he did this to target the research-boosting tradition from Discovery, but the % bonus from the tradition will take some time to catch up to the free research that comes with Mastery of Nature given the low research point income of the early game. Tokryva prefers a setup with stronger late game potential and synergy. However, in the long run a strong player will almost always win a game of Stellaris against the AI. When almost any start will eventually reach dominance in a strong player’s hands, we have to compare choices in terms of how quickly they reach such a state, not WHETHER they will. After all, one can win a game of Stellaris without taking ANY perks! It may feel bad later on to see the useless perk hogging a slot, but early development is critical to quickly and efficiently hitting a strong stride. I think Tokryva is also partially sour on Mastery of Nature because it has been nerfed, such that it feels bad to him taking a perk that once was better.

So who is right? Macsen and Tokryva both make good points, and their differing assessments are rooted more in their metrics for evaluation than in their knowledge of the facts. Macsen is looking to reach a sufficiently winning position quickly, while Tokryva prefers to avoid feelbads (dead later, was better before nerf) and seeks an empire with greater (if unnecessary) potential strength. In terms of tryhard decisionmaking, I think Macsen’s tempo/development-oriented approach has the right of it, but there are merits and good reasons behind both players’ thought processes.