The graphs and data will continue to be updated after publication, but the commentary in this article is as of 5/19/2024 and will not be updated.

In Yugioh it is generally considered advantageous to go first. But how has this advantage changed over time? Also, how big is the advantage?

To answer this question I collected some data on past match history and ran the numbers.

The Result

The rest of the article is just commenting on this graph. If you don't care about me rambling you can stop reading here.

At the end of the article there are some additional graphs including the breakdown by year. I recommend focusing on the by era¹¹ graphs over the yearly graphs, because for the yearly graphs it isn't clear when you are looking at random variation vs a legitimate pattern.

The eras were defined in the following way:

Notes on the data

I think to interpret the result it is good to understand how the data was collected, and what is included.

The data was sourced from written feature matches on https://yugiohblog.konami.com/, and metagame.com feature matches¹, as well as a few YouTube feature matches. If I watched a YouTube feature match for fun, I would enter that match, as well as all the matches for that tournament to avoid bias.

There is a gap in the data around mid 2009 when control of the game shifted from Upper Deck to Konami. There is another gap during Covid when there was no written feature matches.

I read enough of each article to figure out who won the coin flip, who won the match, and who went first. Because the data was manually entered, it is likely there is some error in the data, though I have no reason to believe that this error will bias the result.

All the matches happened at large tournaments; therefore the players are likely to be more skilled than the average player. It is possible the result would be different if a random selection of Yugioh players were selected throughout time. I have no reason to believe the bias would change over time, so the relative win rate should be informative.

In some matches a player won the coin flip but chose to go second. This was particularly common in 2015 when they had just changed the rules so that the first turn player did not draw. It was also believed by many that for Nekroz it was advantageous to go second at the time⁴. In most matches they did not mention who won the coin flip, only who went first. I assumed that choosing to go second is notable enough that if someone did so, it would be mentioned. I also recorded if someone chose to go second in the data. I debated if I wanted to measure the advantage of winning the coin flip or the advantage of going first. After collecting the data I now know that the conclusions drawn would be the same either way. There is more commentary on this in the results.

Matches from side events and dragon duel were not included. This was done because the data was laborious to get through, and I did not believe including the data from these matches would improve the quality of the analysis.

Matches in which one player received a game loss due to a non-gameplay related rule infraction were excluded from the analysis. For example, if there was a decklist error or marked cards the match was excluded, but games where a player had an illegal activation of a card leading to a game loss were included. This was done because it was an in game decision that lead to a player loosing. For example, in 2018 there was a match where a player activated a copy of Sky Striker Mobilize - Engage! with no targets in deck and his opponent chained a face down card revealing hidden information. The game-state was ruled to be irreparable and resulted in a game loss. I believe this was an honest mistake, but because he decided to activate a card and lost as a result of that activation I felt that this was legitimate to include in the analysis. This happened less than 5 times, so the impact is minor. There were a few cases which were ambiguous, but the impact is small.

Only constructed matches³ are included in the data.

If a player was in a position where they would win the duel, but then surrender in order to let their opponent make top cut or go to day two⁵ the player that surrendered is counted as the winner.

The data as well as scripts to generate the exhibits are located here. Please audit the data yourself and contact me if you notice anything else of note.

Results

Why did the win rate change?

It seems the biggest impacts to win rate started around the Synchro era and continued through the XYZ era and then dropped around 2015 with master rule 3.

I think the reason going first became more advantageous around the 5Ds era is that is when it became a good strategy to have a deck that requires setup to perform powerful combos. This meant that the first turn player could set up their combo one turn sooner, and set traps that could effectively prevent their opponent from setting up their combo. For example, a Stardust Dragon backed up by a few good trap cards was very hard to play through, and there was some decks designed to do plays similar to that as efficiently as possible⁶.

This continued until MR3 which eliminated drawing on the first turn. Additionally, around this time monsters got powerful enough that a trap card would be less impactful. After the release of Duelist Alliance it became common for monsters to have floating effects making many traditionally powerful traps weak. I recall activating mirror force against Burning Abyss and by the end of the turn they had gained card advantage. However, it was still rare for monsters to have quick effects, so setting up a first turn board of a lot of monsters was not as valuable as it is today. There were still decks which used a lot of traps like Burning Abyss or Clown Blade, the impact of any one trap was less than prior formats⁷.

As a result, many decks had fairly lackluster first turn plays. For example, the 3rd place deck list at YCS Dallas 2015⁸ only ran two trap cards with no way to search them. This was fairly typical for Nekroz decks which generally preferred to go second. Given this list I may prefer to get a battle phase and an extra card going second. While Nekroz was an extreme case in may ways, I think it does show how the game was shifting at the time.

After this it seems that the going first advantage has been increasing, though it has not returned to pre-MR3 levels of advantage.

The perceived advantage vs the actual advantage

The result is very different than I expected it to be. I would not expect the first turn advantage to be greater in the 5Ds era than the Link era. I think there might be a few factors influencing that perception.

In more recent years it has become more common for the first turn player to set up a powerful end-board and for the second turn player to surrender without playing a single card. There are even decks where the goal is to set up an unbreakable board and win using that advantage, such as Mannadium or Rikka-Sunavalon. While this does not happen most of the time it may have a strong impact on the perception of going first.

Similarly, if you win going second because you happen to draw a card like Sphere Mode, or Evenly Matched you may discount this in your memory as being lucky. For example, Lava Golem was notoriously good against an otherwise very powerful zone-lock Kashtira end-board. If I happen to have Lava Golem in my hand and win as a result, I still walk away feeling like the first turn player had an advantage, because it is easy for me to imagine the game where I did not draw Lava Golem.

Additionally, I think it requires more skill to play going second than going first. When going second you often need to have a good knowledge of how your opponent's deck works, how to optimally interrupt their plays, and deal with the interruptions on their end-board. For example, it might not be obvious to a new player the full impact of Branded Fusion resolving, and that they might want to hold Ash to prevent it, or that Mirrorjade can destroy a monster that is immune to destruction from activated effects. There is also no way to intuit that most of the time when Mirrorjade sends a monster to the graveyard, that monster will have a powerful effect in the end phase. In contrast, when going first it is relatively easy to figure out how to perform your combo, and even if your opponent has a hand-trap your options are still fairly clear. Because all the data was taken from feature matches at large tournaments, this is probably not a factor for most of the games. I suspect that if there was data on matches of less skilled Yugioh players the first turn advantage would be larger. I suspect this influences the perception of the first turn advantage.

More so than anything else, there is a clear difference in strategy going first or second in the modern game which I feel leads to a greater awareness of the first turn advantage.

Going first vs winning the coin flip

I was afraid that for 2015 the high rate of going second would cause a distortion in the results. However, the results are nearly identical if we try to measure the amount of time the coin flip winner won the match. In practice, it seems that winning the coin flip provides a slightly greater advantage than going first, but the difference is very small, and could just be random variation. In 2015 Nekroz was extremely popular, and mostly preferred to go second. This lead to a high blind-second rate in the Arc-v era.

However, even after Nekroz fell out of the metagame blind second decks became more popular than they were in early Yugioh. I think this was because in early Yugioh it was vaguely understood going first was advantageous, but the second turn player's strategy would mostly remain the same. However in modern Yugioh it is possible to form a very different strategy going second.

Closing thoughts

I think this is an optimistic result. The fact that top players consistently top events suggests that Yugioh is a game which rewards skillful play, which is a good thing!

Win rate does not measure everything about the game. A game can have a higher win rate going first, and still be skill-intensive and fun. Chess⁹ and Go¹⁰ both have a bias towards the first turn player, but are widely considered great games.

Other reading

In 2013 Florian Koch wrote and article on the first turn advantage in Magic the Gathering. He uses a very interesting methodology and has a detailed analysis. I highly recommend the article, and you can read it here

Unrelated cool things I noticed

1. Around 2011 they gave the players fun titles in the feature match based on what they were playing. For example, a Synchro plant player was given the title "botanist". One of my favorites was "Monster Hater" for an anti-meta deck²
2. In 2012, they wrote some feature matches from the perspective of Grapha. For example here
3. This deck was really cool: featured here, and here
4. When the Seal of Oricalchos was played in 2013 they changed the text color to green in this match
5. This one guy drew a picture for his feature match request form here and here
6. In 2006 there was someone I felt kind of bad for: Robert Morgan, after going undefeated in swiss, had a series of unfortunate things happen to him. As a result he overslept the start of day 2 and was disqualified. http://kperovic.com/metagame/yugioh5689.html?tabid=33&ArticleId=6083

Other charts

Win rate by year graph

Blind second rate by era