ActiveSG Punggol Sport Centre at 11 Sentul Crescent provides a simple way to understand Information Theory: a sports scoreboard communicates a match state, a booking message identifies a particular date and venue, and a short code can help a receiver detect whether data were copied incorrectly. Information Theory asks how many different messages are possible, how uncertain a receiver is before a message arrives, how efficiently messages can be encoded and how errors can be detected or corrected. The centre began selected operations on 27 October 2026 and has its formal opening scheduled for 22 November 2026. The examples below are fictional. They do not describe the actual Sport Centre scoreboard electronics, MyActiveSG+ security, private booking codes or an installed communications network. Real access is at MyActiveSG+.

MIT’s Information and Entropy course introduces bits, codes, compression, noise, errors, probability and communications, while its advanced Information Theory course develops entropy, Huffman coding, channel capacity and coding theorems. The central concept is that information measures uncertainty and distinguishable possibilities under a stated model, not how interesting a sports story feels. Students should learn the difference between a bit and a byte, a message and the data encoding it, and an error-detecting check and a secret password. A paper scoreboard with four hypothetical states is enough to study the mathematics. This article teaches first principles, worked calculations, ethical communication and a four-week learning plan without attempting to intercept real messages or reveal anyone’s account data.

• Setting: ActiveSG Punggol Sport Centre, 11 Sentul Crescent; any scoreboard or booking scenario is clearly hypothetical.
• Information: distinguishable possibilities described using a defined source and probability model.
• Bit: a binary digit, with possible states zero and one.
• Entropy: expected information in a distribution, H(X)=−Σp(x)log₂p(x).
• Surprisal: the information content of one event, −log₂p, under the probability model.
• Compression: encode information using fewer average bits while retaining the required meaning.
• Prefix codes: codewords designed so one valid word is not the beginning of another.
• Error detection: additional information can reveal some accidental changes.
• Error correction: suitably constructed redundancy can repair specified patterns of errors.
• Communication channel: transmission mechanism with possible noise, loss or limitations.
• Security distinction: error correction does not authenticate a person or encrypt private bookings.
• Safety and privacy: no real OTPs, QR credentials, children’s records, communications interception or live system testing.
• Learning route: definitions, entropy calculations, toy codes, examples, troubleshooting, FAQs and a four-week programme.

Definition lock: information is not identical to meaning

Information Theory gives quantitative ways to reason about signals and uncertainty. A message may be emotionally important but require few binary symbols when the receiver already knows the choices. Conversely, a random-looking identifier may contain many possibilities yet communicate little useful meaning to someone without its context. Students should distinguish the mathematical quantity from personal interest or truthfulness. A sports notification can be factually wrong while being encoded perfectly. The mathematical problem is to represent, transmit or decode symbols accurately under given assumptions. Truth verification and social meaning are additional questions requiring suitable sources and judgement.

A message needs an agreed alphabet

An alphabet is the set of possible symbols used by the model. A toy scoreboard might use symbols W, L, D and P for four fictional states. If sender and receiver do not agree what P means, bits alone cannot solve the semantic confusion. A student should define the alphabet before choosing an encoding. An alphabet of two symbols can be represented by one binary digit under a fixed simple code, while four possibilities require two bits for distinct fixed-length labels. Real scoreboards have more complex information, and no exact Punggol display format is established here.

What one bit means

A bit can take one of two possible values, often zero or one. One bit can distinguish two states under an agreed encoding, such as YES and NO. It is not automatically a one-letter English word. A code designer assigns meanings to the states. If a fictional sport screen needs to show whether a game is active, an invented code might use 0 for idle and 1 for active. A student can draw a two-column truth table to learn the relation. A real operator’s state machine may have many more states; the paper model does not reveal it.

Bits and bytes are different units

A byte ordinarily contains eight bits in modern computing. Eight bits have 2⁸=256 possible binary patterns. That does not mean a byte always represents one readable character: encodings such as UTF-8 may use multiple bytes for a character, and binary values can represent many other things. Students should distinguish stored symbol counts from perceived word length. An imaginary scoreboard record may contain several fields, each with its own encoding. The amount of information cannot be inferred solely from the number of visible letters on a screen. The encoding rules and possible values matter.

Why two bits distinguish four states

With two binary positions, the possible patterns are 00, 01, 10 and 11. Each pattern can label one of four hypothetical scoreboard states. A fixed-length code with one bit would have only two patterns and could not distinguish all four without additional context or a different signalling scheme. This follows from counting possible sequences, not an opinion about screen design. Students can list the patterns and check there are no duplicates. The example also shows why the meaning of a code must be documented: the bits alone do not indicate which state each pattern represents.

Three bits provide eight patterns

For n independent binary positions, there are 2ⁿ possible binary strings. Three bits give eight patterns, four bits give sixteen. A pupil can prove this by observing that each additional bit doubles the existing possibilities. A hypothetical system with five equally valid labels needs at least three fixed-length bits per label, because two bits provide only four distinct codewords. Three bits leave some patterns unused unless extra purposes are assigned. Real systems may choose more bits for compatibility or error detection. Minimal code length is not always the only design objective.

Probability makes some messages more surprising

If a message source generates one symbol more often than another, learning the less common outcome reduces uncertainty by a greater amount. Information Theory uses the event probability to quantify that idea. Students should not confuse a probability assigned in a fictional model with a real SportSG ballot statistic. The probabilities must sum to one and correspond to a defined source. A rare event in a mathematically specified model has higher self-information, but that does not mean it is more valuable or meaningful socially. Information content and emotional importance answer different questions.

Surprisal uses a logarithm

The self-information of an outcome with probability p>0 is I=−log₂(p) bits. A certain event with p=1 has surprisal zero. An event with p=1/2 has one bit, while p=1/4 gives two bits. These examples show the effect of repeated halving. Students should not plug p=0 directly into the ordinary finite logarithm and claim a zero-probability event has a measurable finite surprisal. Mathematical edge cases require careful interpretation. The formula reports surprisal under the stated probability distribution, not a certification of a booking result.

Entropy averages information across a source

Shannon entropy is H(X)=−Σp(x)log₂p(x), summing over outcomes with positive probability using the usual zero-probability convention. It describes the expected information per source symbol under that model. A fair two-outcome source has entropy one bit because both outcomes have probability one half. If the outcome is certain, entropy is zero. Students should write the probabilities and units before calculating. A score display or message source can be mathematically modelled, but actual probabilities must be supported by data if the result is presented as a real measurement.

Worked entropy example: a fair four-state source

Suppose four fictional outcomes each have probability 1/4. Then H=−4×(1/4)log₂(1/4)=−log₂(1/4)=2 bits. A fixed two-bit code is sufficient to distinguish the four states, matching the entropy in this simple equal-probability case. This calculation does not report actual match outcomes or betting probabilities. The fair model is chosen only because it is easy to understand. A student should verify that the probabilities sum to one, then connect the binary code count to the average information result.

Worked entropy example: unequal probabilities

Take a fictional source with probabilities (1/2,1/4,1/4). Entropy is −(1/2)log₂(1/2)−2(1/4)log₂(1/4)=0.5+1=1.5 bits. Three equally likely states would have a different entropy. Students should notice that the formula takes probabilities into account, not only the number of labels. The result is expected information under a model; it is not the length of every individual encoded message. Fixed-length coding for three distinct labels still needs two bits per label, while suitable variable-length coding can average differently.

Maximum entropy depends on the number of possible states

For a discrete alphabet containing m outcomes, entropy is at most log₂m bits, achieved by the uniform distribution under the standard assumptions. This explains why four equally likely states give two bits. A strongly concentrated distribution has lower entropy because the outcome is more predictable. Students should not use “maximum entropy” as a synonym for maximum chaos in every scientific field without defining the model. The same word also appears in Thermodynamics with related but distinct units and interpretations. This lesson concerns discrete information sources and their probability distributions.

Fixed-length and variable-length codes differ

A fixed-length code gives each symbol the same number of bits. For four labels, two bits each may work. A variable-length code can assign shorter words to common symbols and longer ones to rare symbols, reducing average length under suitable distributions. But decoding must remain unambiguous, and overhead or practical constraints may matter. A student should not claim any short code is better simply because it uses fewer bits for one symbol. The relevant measure is the expected length across the source and whether the receiving system can decode every allowed sequence correctly.

Prefix codes prevent ambiguous concatenation

A binary prefix code has no valid codeword that forms the beginning of another codeword. For example, codes A=0, B=10, C=11 form a valid prefix code. A receiver reading 0110 can decode A, C, A by processing codewords in order. The code 0 and 01 together would violate prefix-freeness because 0 prefixes 01. Students can draw a binary tree with leaves at codewords to understand why the arrangement works. This is an educational code, not an actual SportSG scoreboard protocol or encryption system.

Worked average code length

Suppose A occurs with probability 1/2, B with 1/4 and C with 1/4. Using prefix code A=0, B=10 and C=11, the average length is (1/2)×1+(1/4)×2+(1/4)×2=1.5 bits per symbol. This equals the entropy of the model, making the code efficient under these assumptions. A fixed two-bit code for every symbol would average two bits. The example shows why frequent outcomes can receive short codewords. It is not a way to estimate the real size of booking records, which may include validation, storage and security overhead.

Huffman coding builds efficient prefix codes

Huffman coding repeatedly combines low-probability symbols to build a variable-length binary prefix tree. It yields an optimal expected length among binary prefix codes for a given set of symbol probabilities under its assumptions. MIT’s Information and Entropy materials cover source coding and compression. A student can practise with three or four fictional symbols and calculate average length before and after. Huffman coding does not verify whether the message is true or who sent it. Compression and authentication are separate technical objectives. A shorter code is not inherently more secure or suitable for every system.

Compression cannot erase required information

Lossless compression allows the original encoded message to be recovered exactly. If two different source messages were both mapped to one identical output without any other distinguishing information, a receiver could not uniquely restore them. This is a counting constraint. Students should not confuse lossless compression with deleting irrelevant information from a prose article, which may change content but improve readability. In a real system, the code must be specified and decoded correctly. For a toy sports message, every defined state must remain distinguishable after compression if lossless recovery is required.

Lossy compression has a different goal

Lossy compression deliberately permits some difference between original and reconstructed data according to a chosen quality criterion. Images and videos may use lossy techniques to reduce storage. But a booking date or payment amount normally requires exact meaning, not an approximate version. A student should consider which information can tolerate distortion and which must be preserved. A slightly blurred background in a sports photo has different consequences from an incorrectly decoded reservation number. The correct compression method depends on the purpose and acceptable error, not only the largest advertised file-size reduction.

Redundancy can help detect errors

Adding extra symbols may seem inefficient, yet redundancy can reveal whether a message changed accidentally. A simple parity bit is one example. If a fictional set of data bits should contain an even number of ones, an added bit can make that property checkable. A receiver may detect some single-bit changes, depending on the scheme. But parity is not a proof of message authenticity and does not catch every error pattern. Students should explain which errors the code detects and which it can miss. Redundancy is purposeful when it supports reliability, not merely wasted space.

A simple even-parity example

Suppose a toy data word is 1011, containing three ones. Append parity bit 1 so the complete word 10111 contains four ones, an even count. If one bit flips in transmission, the total number of ones becomes odd and the receiver can detect a parity mismatch. If two bits flip, the parity may remain even, so the error can escape this simple check. The lesson concerns accidental alteration under a toy channel. It does not establish the security of an actual account or court confirmation. A parity check has a narrow, precise guarantee.

Error detection and correction are different

An error-detecting code can indicate that something is wrong without telling the receiver which symbol to repair. An error-correcting code adds enough structure to recover certain altered messages under defined conditions. MIT’s Information and Entropy materials treat both noise and correction. A student should not assume every detected error can be corrected. Sometimes a system must ask for retransmission instead. The correct engineering choice depends on channel conditions, cost and required reliability. No actual SportSG communications channel is assessed in this guide.

Hamming distance counts differing positions

Hamming distance between equal-length binary strings is the number of positions at which they differ. The strings 1010 and 1111 differ in positions two and four, so the distance is two. Students can line up the bits and count mismatches. Hamming distance helps analyse how codewords can be distinguished after errors. The minimum distance among a code’s distinct valid words determines some of its error-detection and correction capabilities. This mathematical distance is not a geographical route, a running distance or a measure of personal agreement between people.

Minimum code distance gives a guarantee

A binary code with minimum Hamming distance d can detect up to d−1 arbitrary bit errors in a codeword, under the standard codeword-validity model. It can uniquely correct up to floor((d−1)/2) bit errors using suitable decoding. Thus a code of minimum distance three can correct one error and detect up to two, although a decoder’s behaviour must be interpreted carefully. Students should distinguish code capability from the actual probability of an error. The theorem gives a worst-case pattern guarantee within the model, not a claim that real wireless links always have one-bit errors.

Hamming(7,4) is a classic example

A Hamming(7,4) code encodes four data bits into seven transmitted bits using three parity checks, with minimum distance three in its usual form. It can correct one bit error per valid seven-bit codeword under its specified decoding assumptions. MIT’s errors chapter explains this design as an efficient example of redundancy. Students need not reproduce the full encoder circuit in order to understand the trade-off: four data bits require seven encoded bits in this code. It is not an encryption technique, a Singpass method or evidence of the coding used by any Punggol scoreboard.

Code rate describes useful data fraction

For a block code of length n carrying k information bits, code rate is k/n. Hamming(7,4) has rate 4/7, about 0.571. More redundancy lowers the fraction of raw transmitted bits carrying the original payload but can improve robustness to certain errors. A student should not assume the highest possible rate is always best, because the channel’s noise and required reliability matter. Equally, adding endless redundancy may reduce efficiency without an appropriate benefit. Information Theory studies these trade-offs systematically instead of treating compression and error correction as opposing slogans.

A channel can introduce errors or erasures

A communication channel carries information between sender and receiver. Noise may change symbols, while an erasure means the receiver knows some symbol was not received reliably. These are distinct models with different coding consequences. A student might imagine an electronic scoreboard message transmitted through a toy noisy channel, but the article makes no claim about actual stadium wiring or wireless systems. Define the model: how often symbols change, whether errors are independent, and whether the receiver knows an erasure occurred. A result is meaningful only under the stated channel assumptions.

A binary symmetric channel is a model

A binary symmetric channel independently flips each transmitted bit with probability p under a simplified assumption. If p=0.1, one bit has a ten percent chance of being received incorrectly in that mathematical model. The value is invented, not measured on a public network. MIT’s advanced Information Theory lectures discuss channel models and capacity. A student can simulate a few results on paper, but should not interpret small samples as precise evidence of the underlying probability. Channel models help reason about performance; real networks require appropriate measurement and validation.

Channel capacity is not speed of a sports player

Channel capacity is the maximum reliable information rate under a specified communication model, usually described in bits per channel use or per second depending on the formulation. It does not represent how quickly an athlete runs or how many physical court slots exist. A noisy channel may carry fewer reliable information bits than a noiseless channel with similar raw symbol opportunities. MIT’s advanced Information Theory course develops channel capacity mathematically. Students should define the input alphabet, noise model and output information before discussing capacity. No capacity figure for Punggol’s actual systems is published or inferred here.

Capacity of an ideal binary channel

If a noiseless binary channel can convey one distinct zero-or-one symbol per use, its information capacity is one bit per use under the standard model. If it operates at a known hypothetical rate of one thousand channel uses per second, that corresponds to one thousand bits per second before other practical overheads. This is a simplified example, not a measured venue network. The student should avoid confusing the raw symbol rate with the useful throughput of a real service. Coding overhead, protocols and errors can make actual performance differ from a textbook maximum.

What feedback does in communications

A receiver may confirm that a message was received or request retransmission after a detected error. This is a form of communication feedback. It differs from a control-system feedback loop moving a robot’s motor. Both use observations to influence future actions, but the variables and goals differ. A student can draw a sender, channel, receiver and acknowledgement arrow. An acknowledgement itself may be delayed or lost, which raises further design questions. No actual SportSG network packets need to be captured. The diagram is an abstract way to understand why a sender cannot always assume that transmission implies successful reception.

Repeating a message improves some errors but has a cost

An elementary repetition code might transmit one data bit three times, representing zero as 000 and one as 111. If one bit flips, majority voting recovers the intended bit under the assumption that at most one of the three bits was altered. But the code uses three transmitted bits for one information bit, giving rate 1/3. Multiple errors can defeat it. Students can compare its redundancy and capability with the parity example. It is not a practical authentication scheme or a universal recommendation for a scoreboard. The trade-off between efficiency and reliability depends on the channel.

Packet loss is different from corrupted content

If a message never arrives, the receiver may know only that a timeout or missing response occurred. If it arrives with altered bits, error-detection mechanisms may reveal corruption. A computer network uses several layers to manage such situations, but this guide does not describe the actual software or protocols running MyActiveSG+. A family that sees a spinning payment screen should not assume hacking, lost packets or successful payment from the appearance alone. The authenticated transaction state and official support process determine the practical next step. Information Theory explains possible communication issues, not a diagnosis of one live website event.

Compression and error correction solve different problems

Compression removes predictable redundancy in a source representation so fewer bits are needed on average, while error correction deliberately adds structured redundancy to resist changes during transmission. A communications system may use both at different stages. Students sometimes conclude that adding any extra bits cancels the benefit of compression. In fact, the useful combination depends on source statistics, channel reliability and performance requirements. A toy message can first be represented efficiently and then protected with an error-control code. That conceptual pipeline is separate from encryption and real account security.

Why timestamps carry important information

A sports message saying “Court booked” can be ambiguous when the date, venue and status are missing. More information may reduce the receiver’s uncertainty about which event is meant. Yet adding a timestamp does not guarantee the message is genuine or current. A copied screenshot may show an old but valid historical confirmation. Students should distinguish encoded fields, human interpretation and factual verification. A well-designed message provides the necessary context without exposing personal account secrets. This connects Information Theory to practical communication, but there is no simple entropy formula that certifies whether a public announcement will be understood by every reader.

A booking reference is not automatically public information

A booking QR code or private identifier may contain structured data that the operator uses for validation. Publishing it can create privacy or misuse risks, depending on the system. Students should never collect or reverse-engineer actual booking codes as an Information Theory exercise. Use clearly invented examples, such as a three-bit label from zero to seven, without claiming the real platform uses that encoding. Mathematical understanding of bits and redundancy does not confer permission to inspect private accounts. Secure authentication, authorisation and confidentiality are related disciplines with distinct requirements.

Error correction cannot authenticate a sender

An error-correcting code can help reconstruct a message after specified accidental errors, but it does not prove who created the message or whether the sender is authorised. A fraudster can send a perfectly formed, error-free codeword. Cryptographic authentication uses separate mechanisms and key management. Students should not mistake a visible checksum or parity symbol for proof that a message comes from SportSG. In a real booking decision, independently verified official services remain necessary. The engineering distinction is straightforward: reliability of transmission and trust in the source answer different questions.

A checksum has a limited purpose

A checksum is a computed value derived from data that can help detect some changes. Different checksum methods provide different error-detection capabilities. A simple sum-of-digits example is easy for students but can miss many changes. Cryptographic hashes and message authentication codes are more sophisticated and serve different purposes depending on their use. A student should not call every checksum unbreakable security or claim a visible four-digit code is a digital signature. Learn the specified error model first. All practical examples should use fictional data without attempting to inspect real QR credentials or sensitive records.

Information changes when context changes

Receiving a message provides different information depending on what the receiver already knows. If a family has already agreed to meet only at one venue, a reminder containing the venue name may carry little new uncertainty reduction about that field, though it may still be helpful. If several similar locations are plausible, the exact address can be much more useful. Information Theory formalises some aspects through conditional entropy, but human context is richer than a simple probability table. Students should avoid reporting a numerical entropy for family understanding without measured and defined probabilities. The model and audience must be specified.

Conditional entropy is remaining uncertainty

Conditional entropy H(X|Y) measures the expected uncertainty about X after observing Y under a defined joint probability distribution. If Y tells a receiver exactly which of four states X occurred, H(X|Y)=0. If Y provides no useful information about X in an independent model, the uncertainty remains. A student can draw two fictional cards, one giving the complete state and another showing an unrelated symbol. The concept is useful for asking how much a message reduces uncertainty. It does not measure actual parental confidence or prove that a booking message is factually accurate.

Mutual information measures shared knowledge in a model

Mutual information I(X;Y)=H(X)−H(X|Y) under the standard discrete formulation. It quantifies how much observing one random variable reduces expected uncertainty about the other. For a perfectly copied fair bit, mutual information is one bit. For two independent fair bits, it is zero. Students should distinguish the mathematical quantity from gossip, trust or friendship. A high mutual-information value in an invented signalling model does not show that one actual sport participant understands another person’s intentions. The probability distribution and variable definitions determine the result.

Noise is not only audible sound

In communications, noise refers broadly to unwanted variation or disturbance affecting signals. It need not be a loud noise in a stadium. Electrical interference, transmission imperfections or data corruption can be represented by channel models, while acoustic noise belongs to the physics of sound. The earlier Acoustics guide addresses ears, reverberation and physical sound; this article concerns the reliability and information content of messages. A student should identify the signal and disturbance in the chosen model rather than assume an electronic scoreboard fails whenever the crowd cheers. No venue malfunction is alleged or measured here.

Rate and latency are different performance quantities

Rate describes how much information is transmitted over a duration; latency describes the delay before information reaches a defined point or response completes. A system can have a high raw data rate and still feel slow because of processing, queues or other delays. A student should not assume that doubling the bit rate halves every visible waiting time. In a toy message, a fixed number of bits sent over a channel yields a simple transmission duration, while the full end-to-end delay may include more stages. This distinction helps prevent unsupported claims about the actual MyActiveSG+ checkout speed.

An original fictional scoreboard case

Imagine a paper scoreboard with four states: waiting, in play, paused and finished. Students assign codewords 00,01,10 and 11, then practise translating in both directions. Next they ask what happens if one transmitted bit changes. A changed two-bit pattern may still be another valid state, so no error is necessarily detected. This illustrates why a minimal code designed only for distinct representation is not automatically resilient. The case does not describe any actual ActiveSG scoreboard or match-control protocol. It can be completed with four index cards and no electrical equipment.

An original compression case

A fictional coach sends symbols A, B and C with probabilities one half, one quarter and one quarter. Using a fixed two-bit code would consume two bits per symbol, while prefix code A=0, B=10, C=11 averages 1.5 bits under the specified distribution. Students compute both lengths and compare them with the source’s entropy. The exercise demonstrates how code design responds to frequency. It says nothing about real coach messages, private data or the compression standard used by a sports app. A useful classroom result is the reason for the difference, not just the smaller number.

An original error-detection case

A student writes data word 1011 and appends a bit to make total ones even, producing 10111. A fictional channel flips exactly one bit, so the receiver finds an odd count and detects an error. The student then tries a hypothetical two-bit flip and sees that simple parity may fail. This is a controlled illustration of an algorithm’s limits, not a live test of any public network. The key lesson is that the guarantee must be stated precisely. Error detection cannot tell a person automatically which correct booking date was intended.

An original unreliable-message case

A family receives an unclear fictional note saying “See you at Punggol tomorrow.” The student lists the missing fields: exact location, calendar date, time, sport and whether a booking is confirmed. They rewrite it as a clear invitation using explicitly invented values. The improvement reduces ambiguity for a willing reader but is not a mathematically measured entropy change without a proper probability model. This case connects communication clarity with information concepts while keeping factual verification separate. An attractive or complete-looking message still needs an authentic source when money or access is involved.

A four-week Information Theory learning plan

Week one: binary codes, bit counts, source alphabets and the difference between information and meaning. Week two: surprisal, discrete entropy and fixed versus variable-length codes, including a three-symbol prefix example. Week three: even parity, Hamming distance, error detection and correction through paper strings. Week four: discuss channel noise, acknowledgements, privacy and why authenticated booking information must come from official sources. All exercises use invented data and can be completed on paper. The outcome is understanding the trade-offs among efficiency, reliability and trust, not a claim to have audited any electronic system.

Common Information Theory mistakes

Four possible states do not need four bits in a basic fixed-length binary code. Entropy is not merely the number of symbols and does not measure emotional importance. A variable-length code must remain decodable. Lossless compression differs from lossy simplification. Parity does not detect every two-bit error. Error correction is not encryption or authentication. A private QR code should not be used as a classroom sample. And one failed payment screen does not prove a communications attack. The best habit is to define the source, code, channel and error model before making a quantitative claim.

Frequently asked questions

What is Information Theory? A mathematical discipline studying information, uncertainty, coding, transmission and communication limits.

How many binary patterns do three bits provide? Eight, because 2³=8.

What is entropy? Expected information per source symbol under a specified probability distribution.

Does high entropy mean a story is more interesting? No. Shannon entropy is a mathematical property of the source model.

What is a prefix code? A code where no valid codeword is the prefix of another codeword.

Can parity detect every corrupted message? No. Some multiple-bit changes preserve the same parity.

What is Hamming distance? The number of positions at which two equal-length strings differ.

Does error correction make a message secure? No. Authentication and confidentiality need separate mechanisms.

Can students read real ActiveSG booking QR codes for this lesson? Do not use private credentials or unauthorised account data; fictional codes are sufficient.

Does the guide describe actual Punggol scoreboard equipment? No. All signal and network scenarios are educational inventions.

Official sources and next step

Use MIT 6.050J Information and Entropy, MIT 6.441 Information Theory lectures, and MIT’s Information and Entropy notes for bits, codes, compression, channel noise and reliability concepts. SportSG’s opening event confirms the public venue context; actual booking information remains at MyActiveSG+. Related eduKate guides cover Computing, Cybersecurity and communication through Drama.

Your next step: invent four match states, assign two-bit codes, then add an even-parity check to one message. Explain one mistake the parity detects and one it may miss, without using anyone’s actual account data.

Editorial note: Independent Information Theory education prepared 14 November 2026. No SportSG electronic system, user record, communication link, payment or booking code was inspected or tested.

Continue with the connected guide: ActiveSG Punggol 2027/2028 | Opening Status, Facilities and Visitor Guide. Use its linked index to compare the related questions and choose your next reading.