algonote(en)

There's More Than One Way To Do It

Unity is already profitable

Read Unity's financial report

Runtime Fee will be introduced in Unity

Update: Unity withdrew some of the changes.

It has been announced that Runtime Fee will be introduced to Unity from 2024/01. Unlike the previous billing model, the fee is based on the number of game installations, which has been opposed by developers.

Unity Software Inc was listed on the NYSE in 2020/08. The financial reports are also publicly available. I've read it this time.

investors.unity.com

Unity Overview

Unity was originally founded in Denmark. Closer to it, Spotify was founded in Sweden and Skype was founded in Luxembourg by a Dane and a Swede.

Unity is a game engine, and developing on this platform allows you to deliver games that run on different game consoles, smartphones, etc. The image is that developing on the wrapper is less labor intensive than developing individually. Competitors include Unreal Engine and Godot Engine.

The company's history began when the original game GooBall was not a commercial success. The company pivoted to selling the in-house tools it used to develop the game. Riccitiello, formerly of Electronic Arts, replaced co-founder David Helgason as CEO in 2014.

They are active in M&A. In 2014, they acquired Applifier, Playnomics, and Tsugi, becoming Unity Ads, Unity Analytics, and Unity Cloud Build. In 2019, Vivox, deltaDNA , ChilliConnect, Obvioos, and Artomatix were merged. MLAPI and RestAR were in 2020. Pixyz Software, Parsec, and Weta Digital were in 2021. Ziva Dynamics and ironSource were in 2022.

That's a lot. We need to separate operating income because of the cost of acquisitions.

Unity is already a profitable structure on a quarterly basis

Now let's look at the financial reports.

Annual deficit structure until 2022 is shown in the 2022 Q4 financial reports. It turned profitable in 2022 Q4.

The latest quarterly report also shows an increase in profit.

Unity usage fees will increase by about 20% from 2022/10, so perhaps that is the reason for the increase.

Unity's Business Model

Unity has three revenue streams at the time of the IPO. 54% is from Operating Solutions, consisting of Unity Ads and Unity In-App Purchases. 31% is from Create Solutions, which consists of Unity Engine royalties. 15% is from strategic partnerships and other sources.

Meta was also heavily affected by Apple's restrictions on opt-in of IDFAs (ad identifiers) starting with iOS 14.5 in 2021/04, and there is a possibility that they leveraged others because they were concerned about the future of Unity Ads. In fact, Unity stock crashed 37% in 2022/05 due to IDFAs.

The reporting format has changed since the 2020 IPO. Create Solutions are Create Solutions and Unity Gaming + Strategic Partnerships, which was formerly in Operate Solutions. Grow Solutions include Ads products, Supersonic, Aura, and Luna. Create $508M in the old format, Operate $801. So as a percentage, it hasn't changed from the IPO. I thought there would be more of an IDFA impact.

Hypercasual games are competing with Unity

Here's the complicated part. ironSource and Supersonic were merged in 2015. The merged ironSource was acquired by Unity in 2022. Supersonic is a growth platform, but also publishes its own hyper-casual games. Tall Man Run and Basket Battle are ranked #2 and #5 in the 2022 Hyper Casual Ranking.

According to the 2023Q2 financial reports, the company has already published 90 games.

The newly announced Runtime Fee for Unity is calculated based on the number of installations, so games like hypercasual, which have a low unit cost and are widely installed, seem to be at a disadvantage.

The sales of the games themselves included in Unity's Grow Solutions are not detailed. So it is unclear, but if hyper-casual games are a gold mine as a genre and if hyper-casual games can be easily created using Unity, it is reasonable for the company to make it structurally difficult for other games to be released.

Upselling of existing customers is slowing

The IR material shows the number of customers and Dollar-Based Net Expansion Rate (DBNER: sales growth rate per existing customer).

The number of customers is increasing, but upsells are slowing.

Summary

  • Unity is aggressively acquiring.
  • However, upselling has not been successful for the additional features
  • Reliance on advertising is a business risk due to IDFA restrictions on iOS
  • Recent acquisitions have brought not only game engines but also hyper casual games themselves into the portfolio

So, I thought (delusionally) that it would have been reasonable to make the pricing plan disadvantageous to other hypercasual games by way of an alternative upsell. There is a lot of speculation, so I hope it reads like a conspiracy theory on the back of a flyer.

Comments

The FORM S-1, which is the first part (Japanese IPO format) of the US market, is also available and I would like to read that as well.

www.sec.gov

Not directly related to Unity, but in connection with Scandinavia, The Playlist, based on the story of Spotify's founding, was interesting to see how the founder, who came from an engineering background, worked.

www.netflix.com

Collectible Card Game AI

Notes about the papers of Collectible Card Game AI

Introduction

Game AI competition in IEEE COG 2019 was held. The competitions include Angry Bird level generation, fighting video game like Street Fighter 2, DeepMind hanabi.

Also, Collectible Card Game AI competition was held. One is Legends of Code and Magic, which is simple card game for research. Another one is real digital card game Hearthstone. In Hearthstone page, the list of past papers are available. I've read some papers for curiosity.

Assumption

Legends of Code and Magic and Hearthstone have two tasks.

  1. Create deck from multiple cards

  2. Play good with created deck

Each needs each optimization technique.

There's variation in a. Legends of Code and Magic is selecting cards from limited random set before play(equivalent to Shadowverse 2Pick). Hearthstone AI competition have two tracks. One is play AI using prebuilt deck, the other one is full deck customization & play. The latter one has enough time to consider. The assumption of paper is the same.

Papers

Evolutionary deckbuilding in hearthstone, 2016

Paper about deck creation. With Metastone AI engine, the evolutionary algorithm library μGP is used. To improve confidence, there are 15 tries per deck. The author compared the result with the deck of Tempo Storm, American e-sports team.

Two heroes are scope. They won against the deck created by the professional players. However the simulation of Metastone AI is not perfect.

The computer doesn't play well the deck which human player plays well. Although no condition was assumed, generated deck include multiple same cards.

Automated Playtesting in Collectible Card Games using Evolutionary Algorithms: a Case Study in HearthStone, 2018

Papers from the same authors. After exploring all heroes, they replace the card to similar cost one (+1/-1) to reduce calculations. They call it as Smart Mutation.

In all cases, they outperform human deck. However Smart Mutation doesn't work for some heroes. Most heroes work well and the evolutionary algorithm is stable anyway.

Helping AI to Play Hearthstone using Neural Networks, 2017

AAIA’17 Data Mining Challenge winning solution. The task is to predict win ratio with Neural Network. at the given game state. In this condition, the player doesn't know played cards and remaining cards.

Used network is CNN. The input is normalized to 0-1. Compared with logistic regression, the performance is better thanks to ensemble and tuning. AUC is around 0.01 anyway.

Impression

The assumption of deck selection AI sometimes include play AI. The two tasks are related.

In the both cases, the calculation complexity is important. Play log is the key in n-gram case. Card2Vec may be important.

Future Work

Join competition?

Learn Chinese by Mario Kart

Play notes about Mario Kart Tour

f:id:hiromichinomata:20191024084212j:plain:w200

What is Mario Kart Tour?

Mario Kart Tour is a smartphone game published by Nintendo.

It supports multiple languages including Chinese. In non-Chinese game translation, Chinese sometimes mean only Traditional Chinese. Mario Kart also supports Simplified Chinese. So I played the game for learning.

f:id:hiromichinomata:20191024085045j:plain:w400

Where is text?

The main content of game is obviously race. Characters, item selections are texts. In Mario Kart Tour, you need to choose character, kart and glider before race start. Each has each charasteristics.

The words appeared are described below.

Characters

  • 马里欧 Mario
  • 桃花公主 Princess Peach
  • 耀西 Yoshi
  • 奇诺比奥 Toad
  • 酷霸王 Bowser
  • 森喜刚 Donkey Kong
  • 迪迪刚 Diddy Kong
  • 嘿呵 Shy Guy
  • 慢慢亀 Koopa Troopa

It seems translation come from sound. Koopa Troopa comes from visual image.

Item

  • 赛道 Track
  • 角色 Character
  • 杯赛 Cup match
  • 宝贝 Baby
  • 肥皂 Soap
  • 蘑菇 Mushroom
  • 冲刺 Spurt
  • 提升 Lift up
  • 喇叭 Trumpet

Baby is treasure.

Impression

As expected, it's race game. Texts are minor part.

Future Work

Novel game?

How to learn Chinese By Ring Fit Adventure

Exercise with Chinese

f:id:hiromichinomata:20200225231708j:plain:w200

What is Ring Fit Adventure?

Ring Fit Adventure is the exercise game on Nintendo Switch. In addition to the Switch itself, ring controller is needed for play.

One Switch controller(Joy-Con) inserts to the ring. The other side inserts to the band on the thigh. The movement of the body is tracked by the combination of the sensors on the two controllers.

Ring-side Joy-Con can measure heart beat ratio too.

Gakky's CM is famous.

Exercise Chinese

This game is translated to many languages including Chinese. Not only subtitles, also speech instructions are supported.

Playing the game in Chinese is the main part of this article.

Although I played in Simplified Chinese, the vocabulary below is Traditional Chinese due to the readiness of extraction. English and Japanese translations are also described.

Muscle Training

  • Back Press 背面推壓 ウシロプッシュ
  • Overhead Press 高舉雙臂推入 バンザイプッシュ
  • Front Press 向下推壓 サゲテプッシュ
  • Bow Pull 圓環箭 リングアロー
  • Shoulder Press 肩部推壓 カタニプッシュ
  • Tricep Kickback 三頭肌 トライセプス
  • Squat 深蹲 スクワット
  • Wide Squat 寬深蹲 ワイドスクワット
  • Overhead Squat 高舉雙臂深蹲 バンザイスクワット
  • Overhead Side Bend 高舉雙臂側彎 バンザイサイドベンド
  • Overhead Lunge Twist 高舉雙臂扭動 バンザイツイスト
  • Pendulum Bend 俯身划船 ベントオーバー
  • Overhead Bend 高舉雙臂晨式 バンザイモーニング
  • Thigh Press 大腿推壓 モモデプッシュ
  • Knee-to-Chest 抱膝式 ニートゥーチェスト
  • Seated Forward Press 往前推壓 マエニプッシュ
  • Hip Lift 提臀 ヒップリフト
  • Plank 平板支撐 プランク
  • Leg Raise 提腿 レッグレイズ
  • Open & Close Leg Raise 腿部開合 アシパカパカ

Rhythm

  • Standing Twist 甩手 スワイショウ
  • Overhead Hip Shake 高舉雙臂扭腰 バンザイコシフリ
  • Overhead Arm Twist 扭動手臂 アームツイスト
  • Overhead Arm Spin 轉轉手臂 グルグルアーム
  • Russian Twist 俄式扭腰 ロシアンツイスト
  • Flutter Kick 踢腿 バタバタレッグ
  • Seated Ring Raise 抬放Ring-Con リングアゲサゲ
  • Mountain Climber 登山式 マウンテンクライマー
  • Leg Scissors 剪刀式 ハサミレッグ
  • Knee Lift 抬抬大腿 モモアゲアゲ
  • Side Step 踏步 ステップアップ
  • Ring Raise Combo 連續抬放 アゲサゲコンボ
  • Knee-Lift Combo 連續抬腿 モモアゲコンボ

Yoga

  • Chair Pose 椅子姿勢 椅子のポーズ
  • Tree Pose 立木姿勢 立木のポーズ
  • Hinge Pose 合頁姿勢 チョウツガイのポーズ
  • Revolved Crescent Lunge Pose 扭轉側角姿勢 ねじり体側のポーズ
  • Warrior I Pose 英雄1的姿勢 英雄1のポーズ
  • Warrior II Pose 英雄2的姿勢 英雄2のポーズ
  • Warrior III Pose 英雄3的姿勢 英雄3のポーズ
  • Fan Pose 扇形姿勢 扇のポーズ
  • Boat Pose 船式 船のポーズ
  • Standing Forward Fold 摺疊軀體姿勢 折りたたむポーズ

Impression

The exercise is very hard. Before learning language, it seems I need more basic training...