Original on Beyond the Duel
(Thanks for Fuyuki’s translation and bring this article to English society)
Originally found on Weibo. Reprinted and translated with permission. Original author: 影融
Why is playing online so different from playing offline?
Hello everyone, Shadowmeld (影融) here.
This article is different from the typical ‘deck insight’ that I usually write. Instead, this is more of a summation of what I’ve been musing about lately. Furthermore, the writing process itself is an opportunity to distill my experiences and reach a solid conclusion.
Table of Contents
- 2、Online/Offline Metagame Differences
- 3、Online Meta: The Birth of the Toolbox
- 4、The Toolbox Dilemma
- 5、Resolving the Dilemma
It is a widely observable phenomenon that despite sharing the same banlist, there is a large divergence between decks that are popular online, at locals, and at larger tournaments. Dear reader, have you considered why certain decks that do well at tournaments are unpopular online? And indeed, the reverse is often true as well.
Before authoring the article, I wager that many of you would say, “It’s just different playing online and in person.” Or perhaps “Japan sucks at the game”. However, there is rarely any consideration for the factors at play behind the phenomenon. I have pondered this question for a while now, and in the ensuing text, I intend to put to paper my thoughts on the subject – and to fundamentally illustrate the reason for the divide between online and offline.
 TN: It is commonly the case that very odd decks win smaller tournaments in Japan, but this is not sufficient explanation.
First, I will begin by defining terms and listing my sources of data for ease of later discussion.
- Online Meta: The competitive ladder @ mycard.moe, which includes games played at DP 1400+ and YGOPro All-Stars tournament results. Data provided courtesy of the same.
- Locals Meta: The metagame that can be found at various local game stores. The winning decklists can often be found on the social media of these LGSes, and the Yu-Gi-Oh! Info Review Weibo account does metagame reviews every week from Ranking Tournaments as well. All locals-level metagame data is included here, including those of Duelist Cups and Japanese Qualifiers.
- Tournament Meta: More accurately ‘Chinese large-scale tournament meta’, but abbreviated for reference’s sake. Specifically, China City Tour. All data collated by the Yu-Gi-Oh! Info Review Weibo account. Because only a fraction of topping decklists are made available to the public, the social media of top players is a good source of supplementary information to learn more about the decks that do well at these events.
- Japanese Meta: Including only larger events in Japan, such as Nextplay Cup. Metagame breakdowns are a good snapshot of the Japanese tournament meta.
Now that we understand what we are talking about, let’s proceed.
The following are some metagame breakdowns.
As is evident from the above, the metagames reflected in these breakdowns are vastly different. It cannot be arbitrarily said that one metagame is somehow at a higher level than the others, e.g., that Japan tournaments are more conclusive than Chinese ones. However, it is a matter of fact that a deck most suited to a certain metagame may not do as well when transplanted to another. If our objective is to craft the deck most suited to a target metagame, we must be mindful not to generalize. What does well in one meta may not in another.
Considering the MyCard ladder usage statistics and the online Yu-Gi-Oh! All-Stars invitational meta breakdown, we note that players in the 2022.1 OCG metagame tend to favor Marincess, Floowandereeze, Drytron, and Adamancipator. This is drastically different from the offline metagame, where we know for a fact that the most popular deck is Adventurer Branded Despia.
What are the factors at play here that have caused this dichotomy?
3.1 Characteristics of Online Play
Playing online has three main characteristics. Better random shuffling, more frequent misplays, and a wider range in opponent quality.
Whether we like to admit it or not, shuffling decks offline cannot come close to a true random distribution. Although the algorithms responsible for shuffling online are technically pseudorandom, they produce results that are much closer to ‘true randomness’.
I think, that if a starting hand were to be scored, although the average score for the same deck may be the same online or offline, it is more likely for online hands to give extreme results on either side of the curve, meaning fewer hands are closer to the average. In contrast, because of the imperfections of shuffling offline, it is far more unlikely for us to draw truly amazing or truly terrible hands. In other words, more hands will be close to the average.
As the graph below shows, if we model the score of a hand using a normal distribution, the blue curve might represent the distribution of offline hands (more results closer to average), while the red, online hands (more results toward the extrema).
The blue curve is more prominent in the center; the red curve has more extremes
The second point is the possibility of misplaying. Perhaps you may have had this kind of experience playing Yu-Gi-Oh! Online as well. It is much easier to make very simple mistakes that have disproportionately disastrous consequences. Misclicking, not turning on full control over chaining options and therefore missing your desired timing, and other similar errors are rampant in online Yu-Gi-Oh. The consequence of this is that many players will choose comparatively more fault-tolerant decks, i.e., decks that are simpler and harder to make serious mistakes with.
At the same time, YGOPro has a 3-minute turn limit. If you play a deck that has a large number of steps in its combos, you will not have any time to think about what you are doing. This is an indirect encouragement of decks that have so-called brainless combos, with fewer steps and very little thought required. Simpler, midrange decks with a focus on resource generation and grinding become more favorable in such a metagame.
Finally, there is the factor of opponent skill at play. In the online environment, it is possible to match against players of all skill levels. More importantly, there is an abundance of weaker, more unskilled players. Consequently, using ‘cards that are stronger against weaker players’ becomes an encouraged action.
The usage of Trap-heavy decks is a very good example of this. In offline metagames, Trap-heavy decks are generally nowhere to be found, but they are quite commonplace online, especially at the lower ranks. This is because weaker players are usually not very good at baiting out and playing around and through backrow disruption. As such, Trap Cards become more valuable overall.
Based on these factors, the MyCard Competitive Ladder makes these demands of its players:
- The ability to play with weaker hands.
- Higher fault tolerance. Not misclicking is ideal, but the consequence should not be severe even if one does misclick. Lower time-to-think is better.
- Can easily beat weaker players. There are far more weak players online compared to strong ones. However, not much more DP is gained from winning against the latter versus the former.
3.2 Evolution of the Commons: A Greedier Strategy
Based on the above demands, players have realized that a high-risk, high-reward strategy that is more opportunistic is best suited to taking on the MyCard metagame. (Call it a greedy strategy for short.)
Greedy strategies generate good returns. When there are powerful dominant decks in the meta, such as Branded Despia for this one, mirror matches become quite likely. However, mirror matches are already error-prone affairs, and the aforementioned misclicking, missing timing, and the 3-minute time limit are all additional to those that can be made offline. In light of these issues, the addition of greedy tech choices such as Super Polymerization can drastically bolster the winrate of the mirror.
More importantly, greedy strategies often go unpunished against weaker players.
As mentioned before, targeted teching is a classic implementation of a greedy strategy (at least for Yu-Gi-Oh). By including these tech cards, though, the possibility arises that they become useless when drawn at an inopportune time – thus slightly lowering the quality of the average hand. And yet, when the average opponent is weak, either due to skill or deck type, the quality of their hand as it translates to what they can put onto the board will not result in punishment for the player with the greedy strategy.
For example, we draw Nibiru, the Primal Being. This is a card that is useless in the current matchup. However, our opponent is playing Rai-Oh Boarder Beatdown, which is not a deck that will punish us for a single pseudo-Garnet in our hand. Considering our opponent’s strange deck choice, they may even be a new player, which means they will have no opportunity to punish us whatsoever. However, if this example occurred offline and our opponent was playing the very common Branded Despia, we would definitely be put at a considerable disadvantage by virtue of having a useless Nibiru in our hand.
Another example of greed would be to play a deck that has a larger difference in winrate going first or second. When going first, the greedy player has a deck that is sufficiently strong enough to set up backrow or combo off for an extended amount of time, giving him an easy victory.
In an offline environment, these decks’ weakness of going second means that they have a hard time getting past opposing counter boards and/or disruption, which is the failure state (getting punished). However, playing online may not yield the same result. Even if the opponent is playing a midrange deck, due to the existence of weaker players on the online ladder, the counters and disruption from said midrange deck may not be sufficiently effective at stopping the greedy deck’s plays.
Take for example a match between Branded Despia and Wyrm Goodstuff. With the same one-Albaz board of Mirrorjade the Iceblade Dragon and Predaplant Dragostapelia, the Branded Despia player needs to use their disruption at the most crucial moments in the Wyrm player’s combo in order to stop him from setting up that turn for good. Because of the wider range of player skills, the chances of this happening correctly online are considerably lower. When an online opponent uses their disruption incorrectly and fails to stop a greedy player’s turn, that too is an avoidance of punishment for the latter.
Let us summarize. The greedy strategy is a high-risk, high-reward strategy that seeks to maximize winrates against the online field. The considerable risk involved with the strategy means that two outcomes are possible: successful greed and failure to greed. Against strong players, taking the risk successfully means victory, while failing means punishment. Against weaker players, even when the risk fails, there is a chance that the risk-taking player will not be punished. It is precisely because of this last possible outcome that the detriment of failure is dampened, resulting in an overall increase in the expected value of greed – which naturally encourages more risk-takers.
What are these risk-takers playing? Two or three years ago they were playing Trickstar, Trap-heavy Orcust, Altergeist, and Subterror. One or two years ago, Trap-heavy Code Talker and Eldlich. Nowadays, they are playing Floowandereeze, Drytron, Adamancipator, Dinomorphia. None of these decks were heavily used in offline metagames of their time, but they have achieved tremendous popularity online.
As shown below, MyCard’s top ten decks by usage statistics. Most of the decks here have the characteristics mentioned above.
3.3 The Spike’s Final Form: The Toolbox
In a departure from the last section, we will speak about how Spikes (the subset of strong players, rather than all players) approach the online metagame.
When a plethora of online players become greedy risk-takers, other players – especially those who are already competitive offline – will inevitably make adaptations accordingly, which leads to the birth of:
Competitive players want to win as many games as they possibly can. However, when faced with greedy decks, it is difficult for them to pilot conservative, resource-focused decks to victory (conservative in this case means a deck that does not greed or try to take overt risks). For example, Adventurer Branded Despia, a deck commonly seen in Japanese tournaments.
While it is true that conservative decks are adept at winning grind games, the increase in risk-takers in the metagame means that the value of doing so goes down: if the opponent’s greedy choices prove successful, they will have already obtained a significant advantage that is difficult to incrementally overcome; and if the opponent fails, they will already have been punished, which means the plentiful resources generated by a conservative deck become a bit overkill in achieving victory.
To make this easier to understand, let us consider some fairly extreme examples: the Sky Striker mirror, from several formats ago. If the opponent is a risk-taker, e.g., they play Shared Ride and White Howling in the main deck, our winrate against them will certainly be impacted despite our skill; on the other hand, if they draw into multiple copies of Nibiru – a card teched against explosive combo decks and not useful in the mirror – they are already at a disadvantage, which means we don’t need so many advantage or grind cards to win. Both these cases come from personal experience.
Is it really true that I have nothing to do with whether I win or lose? Although it is hard to admit this, when a conservative, resource-focused deck faces a greedy, risk-taking deck, the result of the match will be considerably affected by whether the latter successfully greeds against the former.Since that is the case, it does not matter so much whether we are good at squeezing resources out of our own decks (as is reflected in choices made in the deckbuilding stage).
This forces a change in conservative decks that find success in offline metagames. Resource engines can be slimmed down, because they do not play as much of a factor in deciding the outcome of a game. The opened slots can then be used for tech cards.
When there is a significant number of risk-takers in the metagame, it is necessary to play greedy cards yourself targeted against greedy players in order to maintain winrate. This is because if Spikes do not tech against greedy players (set 5 pass / first-turn solitaire), their winrate will be negatively affected when going second, which is something that they simply cannot tolerate.
The rise of the tech arms race means an increased number of tech cards are being played in the main as well. Cards like Nibiru, the Primal Being and Harpie’s Feather Duster have a much higher occurrence rate there, compared to offline tournament toppers. There are no concrete statistics to support this, but anecdotal evidence from many players corroborates it.
And yet, as you may have noticed already, tech cards against greedy players are a greedy choice as well. As explained above, risk-taking behaviour encourages proliferation of the phenomenon in the metagame. As time goes on, risk-taking greediness comes to dominate the metagame.
If it is necessary to take risks regardless, why not tech offensively instead of defensively against your opponents? To go a step further, why not tech cards against the entire field?
This is the birth of ‘the Toolbox’.
To illustrate it more aptly, let us observe two very powerful decks in the online metagame.
2022.1 MyCard Ladder Rank 1
This deck differs from the typical Floowandereeze build by not playing Barrier Statue of the Stormwinds. Instead, it plays a higher number of high-level monsters for the mirror. Harpie’s Feather Duster is played in the main to fill the deficiency in backrow removal. Instead of the typical Maxx ‘C’ and Vanity’s Emptiness in the main, Mist Valley Apex Avian is used for its ability to negate backrow. The Winged Dragon of Ra – Sphere Mode is played in the side as well. These tech choices are excessive when considering the overall field.
YGOPro All Stars Champion
This build also plays Duster in the main deck. The inclusion of Nibiru allows it some ability to fend off combo decks. Droll & Lock Bird and Ghost Belle and Haunted Mansion are similar techs against slightly different breeds of decks. Vanity’s Emptiness and Monster Reborn are included for additional utility (explosive combo decks like Adamancipator rarely play Vanity at all).
This is my summation of Toolbox-style decks.
Toolbox-style decks play greedy cards in the main deck. These cards are not part of the deck’s usual resource engine, but rather cards that fulfill some sort of function – in essence, they are tools to be used when the time is right. When facing a greedy opponent, a Toolbox gets into an arms race against said opponent to see whose tools are more appropriate for the job; at that point, the outcome of the match no longer has much to do with which deck purely generates more resources.
In a metagame with a large number of greedy players, the Toolbox is born. In my opinion, the Toolbox is the evolution of the greedy deck, and a reasonable choice in a developed, risky metagame. It retains the basic deck type’s resource engines, but plays techs and anti-techs on top of that deck. Including so many techs means that there is almost always one tool that’s right for the job – allowing the pilot of this kind of deck to play evenly against any risk-taker, in theory.
The Toolbox is also a deckbuilding philosophy. When a metagame has great diversity in deck types, there are inevitably plenty of greedy decks that need to be teched against. Because other people are playing techs, anti-tech techs are needed to stay relevant against said techs. Now that there are so many techs in the deck, it’s hard to win a grind game anymore, so why not play even more techs? In the end, the plethora of tech choices in the deck form a Toolbox.
From our conclusions above, it seems that the Toolbox is perfect. But is that really the case?
Let us return to the question at the beginning of the article. Why do online and offline metagames differ so much?
We notice that the Toolbox, which is the most favorable strategy to take when online, is no longer effective in paper tournaments. The decks that top in these situations tend to be conservative decks most of the time, and the decks that do well online do not tend to perform well offline. Why is that the case?
The fundamental reason is that more players are risk-averse in the offline metagame. One of the most realistic reasons is that it costs money to travel to a tournament, which is a significant detriment to people’s willingness to take risks; whereas being punished for greed online is merely a loss of a few ladder points. Of course, there are other reasons as well, like misclicking and missing timing generally not being a factor offline… but these have been discussed already, so I will not elaborate on them further here.
The Toolbox is effective in a metagame where many players are risk-takers, and choose greedy decks. However, when the Toolbox – with its huge diversity of tech choices – faces an individual opponent, there are inevitably one or several cards in its arsenal that are useless in the matchup. This occurrence is already a punishment, since the Toolbox player is playing a few cards down (either in the deck or in the hand); when the advantage generated by favorable techs do not outweigh this detriment, we notice that the Toolbox will lose to the conservative deck.
This is especially common in the recent metagame. Because of the prevalence of the Adventurer engine, it is difficult to tech against the most popular deck in the meta; this has been the case for two major formats now (the two of them being DPE Adventurer PK, and Adventurer Branded Despia). This is also why I think Adventurer should be banned out in the April list.
Because of the difficulty in teching against the most popular deck, the natural inclination is to include even more techs against it, or perhaps play a deck with a favorable matchup; but aggravating this behaviour makes it more difficult to win against other decks in the metagame.
If there are a significant number of techs against the strongest deck, the player ends up drawing more dead cards against other decks. In order to increase one’s winrate against different decks, more and different techs are then needed, damaging the effectiveness of the deck’s main engine(s). This means that some games which are favorable by matchup alone no longer become so favored, which leads to a further need to tweak… and so on and so forth. After a certain point, greedy techs are insufficient in winning a game alone.
This is the Toolbox Dilemma.
For this section, the reader may consult the winning decklists for the last several tournaments. No lists will be included in the article.
How shall we free ourselves from the dilemma? The pursuit of the hottest and best techs?
The answer actually lies in plain sight already.
In my opinion, resolving the dilemma involves changing one’s mentality and pursuing a different, more defensive strategy. Especially in the main deck, it is best to give up aggressive techs in favor of more reasonable conservative and defensive choices. Without taking risks, there is no punishment. It is far easier to ward off the attacks of other risk-takers, forgoing the winrate lost to their successful attempts at greed and focusing on punishing their failures to do so.
However, if one plays only conservative engine cards in the main deck and also plays the most popular deck in the format, some players will take on weak risk by teching only against the most popular deck and no one else. It is impossible to punish these weak risk-takers because they have not taken any other risks to capitalize on, which is ultimately detrimental to the conservative strategy. For this, I propose several potential countermeasures:
- Play a denser main deck. Play the most popular deck and even more starters for the engine. This is, however, difficult to practically achieve because usually starters are maxed anyway. Playing more Pot cards and a slimmer main deck can help toward this goal.
- Play a rogue deck. Take advantage of an information gap by bringing something that people do not expect – and therefore will not tech against. Obviously, don’t play a deck that gets beaten by the same techs as the most popular deck.
- Play a deck that counters the most popular deck. Basic rock-paper-scissors that has been present in many formats. There are minute differences against different builds that may lead to favorable and unfavorable matchups as well. This is a test of one’s meta knowledge and deckbuilding skills.
(Translator’s note: Honestly, these are all pretty obvious strategies, but hey.)
The Toolbox philosophy is not completely useless in tournament environments, either. I think that the main and side decks should, when considered as a whole, include tools to play against any kind of deck. If a conservative main deck is necessary, then at least include appropriate tools in the side to deal with a variety of situations.
This is directly tied to playing more powerful, focused techs in the side. Reflecting the Toolbox philosophy, these techs are aimed at aggressively attacking the decks they are intended to be used against. It is best to play tools that bluntly win single matchups, rather than ones that are weakly helpful against multiple but never end up being deciding factors.
Don’t take risks in the main. Choose the right deck. Play focused techs in the side.
This article was a bit stream-of-consciousness, and some parts of it may appear counterintuitive, but these really are conclusions I have come to after organizing my thoughts recently. It was not easy for me to pen them down, but I wanted to try doing so anyway; I did not expect to produce something that I could be satisfied with to a reasonable degree.
Now that I have written it down, I hope that it is helpful to all of you who have read up to this point.
All decks discussed in the article are reprinted for discussion and analysis purposes only. Thanks to all the players who built the decks in question. All images included have sources included, and may be deleted pending any copyright violation.