An Empty Spreadsheet in Penang: Reading Badminton for a Living and the Right to Stay Silent

2:47 a.m., Penang. The ceiling fan turns, the coffee went cold a while ago, and on screen a spreadsheet has opened with exactly one value: 0.

Not 0 goals, 0 points, 0 rallies. 0 valid rows. The columns are there, the rows are empty. No rally length, no serve-point distribution, no shuttle flight time, no footwork speed. My trade has a short name for this state: insufficient information.

The phone buzzes. A regular reader in Kuala Lumpur: “Can you read this match for me, the rally over-under is high.” I look at the empty sheet, then at the message. The honest answer is the one it took me years to dare to say out loud: there is nothing to read.

When there is nothing to read, the only truthful conclusion is no conclusion.

That is the first lesson, and the hardest one. Outsiders assume betting analysis is a talking job. Talk a lot, talk fast, talk certain. The reality runs the other way: it is a job of refusing to talk, and speaking only when there is something to say. Penang is where I buried a part of my innocence; since then I have dug data the way others dig graves.

To form a view on a men's singles badminton match, I need at least four layers of data. The first is average rally length and its distribution — not a bare mean, but the shape of the distribution, because a match of nothing but short rallies and a match where a third of the rallies pass 20 shots are two different sports. The second is the win correlation between serving points and receiving points. The third is error type: shuttle out, shuttle into the net, or an unforced error from choosing the wrong line. The fourth is set structure — who wins the back half of each set, who wins from 18 points onward.

Miss one layer and the conclusion tilts toward the layers that remain.

With rally length but no error type, I will talk about fitness when the problem is shot selection. With error type but no point distribution, I will talk about technique when the problem is the end of a set. With point distribution but no serve correlation, I will tell a psychological story when the problem is serve tactics.

Malaysia is the densest badminton market in Southeast Asia. The BWF World Tour passes through every year, the Malaysia Open is a fixed stop in the Super 1000 tier, and the betting audience here reads badminton the way others read a price board. The country's men's doubles foundation has one landmark worth citing: Aaron Chia and Soh Wooi Yik won the world title in 2026 in Tokyo, as recorded by the Badminton World Federation. Men's singles has Lee Zii Jia. Women's doubles has Pearly Tan and Thinaah Muralitharan, Malaysia's top pair for years.

In that ecosystem, an analyst with no data is a useless analyst. But an analyst with an empty sheet who still posts a pick is worse than useless — he is not merely wrong, he makes readers believe the wrong thing.

Here I should confess a professional habit. Every time I receive a dataset, I check it three times. First, provenance: which collection system produced it, who entered it. Second, consistency: does the total points in the file match the published score, does the total rally count match the match duration. Third, reproducibility: take a small sample, count it by eye, compare with the file.

Those three checks are not ritual. They are the only thing separating an analyst from a storyteller.

In 2026, while working as an analyst for a newly launched television channel in Malaysia, I published self-collected xG data for the match between Pulau Pinang and Johor Darul Ta'zim: the home side generated 2.8 xG yet lost 0-2. I was heavily criticised for saying the losing team had created the better chances. A week later the Pulau Pinang head coach lost his job, and the team won four straight under the assistant. The data stood on my side, but the lesson I kept was not that I had been right. The lesson was that a correct number can still be used wrongly, and whoever uses it owns the way it is arranged.

I do not trust any statistic that cannot be arranged — in the professional sense: arranging means reordering the story the raw number is hiding. If I cannot show where it sits in the chain of cause, it is not data, it is decoration.

Then came the pandemic season. When the Bundesliga 2026-2026 restarted in empty stadiums, I set out to collect 145 matches and measure the so-called disappearing home advantage. The result: home win rate fell from 43 percent to 31 percent, while the over rate rose by roughly 12 percent. I published it and Western analysts attacked it for a small sample. The right response is not to defend the conclusion but to widen the sample. I tracked 98 more matches in Hungary and Portugal. Only later did major outlets cite the work as a unique study of pandemic football.

An empty stadium is like a prayer rug; the odds tremble along every nerve. When the crowd noise vanishes, what vanishes with it is not emotion but a variable in the model. And when a variable vanishes, every model still running is wrong to some degree.

That is why I told the reader in Kuala Lumpur that night I had nothing to read. An empty sheet means every model of mine is missing variables. A model missing variables does not produce a forecast; it produces bias dressed up in numbers.

This trade has three ways to fill a data gap, and all three are traps.

The first is borrowing old numbers. Take last season's statistics, or the nearest comparable event, and attach them to the match in front of you. This is more dangerous in badminton than in football, because conditions shift fast: shuttle speed depends on arena temperature and humidity, and the same player can produce two completely different matches in one week in two different halls. An old number is not a wrong number. It is the number of a different match.

The second is turning correlation into causation. A player wins repeatedly in a fast hall, and the conclusion becomes that he has upgraded to an attacking game. Possibly true. It is equally possible he simply met three opponents with below-average defence. Without opponent data, the two hypotheses cannot be separated.

The third, the most common and hardest to spot, is telling a story so good that the reader forgets the data was empty. A beautiful sentence about an athlete's maturity carries more weight than a blank table. But that weight is the weight of prose, not of evidence.

Three months living with the World Cup taught me: money never runs in a straight line. It runs on news, on hours, on liquidity. In badminton the in-play money is steeper still, because each rally lasts a few dozen seconds and each set can flip inside four points. Get one read wrong and the reader loses money faster than it takes me to finish the sentence.

The counterintuitive angle sits here: in this trade, silence is a conclusion, and it has value.

The whole ecosystem is built to reward talkers. Bookmakers need money to move, platforms need content, readers need an answer before the first serve. Inside that structure, the phrase insufficient information gets read as weak, as underconfident, as having nothing to sell. Look closer, though, and an empty dataset is a valuable signal: it says nobody in the collection chain did their job, that this match is being priced on feeling, and that most participants are betting on a memory of the previous match.

Knowing what you do not know is a position. It does not pay directly, but it stops you from betting on what you mistake for knowledge.

There is a limit, of course. Saying insufficient information three times a week is dodging work. The difference between humility and laziness is this: the humble analyst went and fetched the data and found it absent; the lazy one never went. My own rule: before declaring a dataset empty, prove you tried at least three independent sources.

In the coming rounds, when the BWF World Tour returns to Southeast Asia, I will track one specific signal: the share of rallies passing 20 shots in the second game of three-game matches. If that share rises steadily in a group of players whose rest windows are shorter than everyone else's, that is the trace of something the scoreboard never shows.

Tonight, though, the sheet stays empty. I answered the reader in Kuala Lumpur with one line, then shut the laptop: when the data goes to zero, my writing has to go to zero too.