The SIG Crash Course

Options are the beating heart of SIG. The firm was built on options market making, and it remains the foundation of everything they do. To understand SIG, you must first understand options — not just the mechanics, but the philosophy behind them. Options don't just let you bet on direction; they let you bet on how much something will move, when it will move, and the uncertainty itself.

The Fundamentals: What Are Options?

An option is a contract that gives the holder the right, but not the obligation, to buy or sell an underlying asset at a specified price (the strike price) on or before a specified date (the expiration date). For this right, the buyer pays a premium to the seller.

Call Options

A call option gives the holder the right to buy the underlying asset. You buy calls when you think the price will go up. If the stock is at $100 and you hold a $105 call, you profit if the stock goes above $105 (plus whatever premium you paid).

Put Options

A put option gives the holder the right to sell the underlying asset. You buy puts when you think the price will go down, or when you want insurance. If the stock is at $100 and you hold a $95 put, you can sell at $95 even if the stock falls to $50.

The Anatomy of an Option

Component	Meaning	Example
Underlying	The asset the option references	AAPL stock
Strike Price	The price at which you can buy/sell	$150
Expiration	When the option expires	March 21, 2026
Premium	The price you pay for the option	$3.50 per share
Type	Call or Put	Call
Style	American (exercise anytime) or European (exercise at expiry only)	American

Moneyness

Moneyness describes the relationship between the strike price and the current stock price. It's the first thing a trader assesses:

• In-the-money (ITM): The option has intrinsic value. For calls: stock price > strike. For puts: stock price < strike.
• At-the-money (ATM): Strike ≈ current price. These are the most actively traded.
• Out-of-the-money (OTM): No intrinsic value. For calls: stock price < strike. For puts: stock price > strike.

Intrinsic vs Extrinsic Value

Every option's price (premium) decomposes into two parts:

Intrinsic value is the amount by which the option is in-the-money. A $100 call when the stock is at $108 has $8 of intrinsic value. Intrinsic value can never be negative — if the option is OTM, intrinsic value is zero.

Extrinsic value (often called time value) is everything else — the premium you pay for the possibility that the option might become more valuable before expiration. It represents time, volatility, and uncertainty. Extrinsic value always erodes to zero at expiration.

The Greeks: Measuring Risk in Every Dimension

The "Greeks" are partial derivatives that measure how an option's price responds to changes in various factors. If options are the language of risk, the Greeks are its grammar. At SIG, traders don't just know the Greeks — they think in Greeks.

Delta (Δ) — Directional Sensitivity

Delta measures how much the option price changes for a $1 move in the underlying. A delta of 0.60 means the option gains approximately $0.60 for every $1 the stock rises.

• Calls: Delta ranges from 0 to +1. Deep ITM calls approach 1; deep OTM calls approach 0.
• Puts: Delta ranges from -1 to 0. Deep ITM puts approach -1.
• ATM options: Delta ≈ ±0.50 (roughly a coin flip on whether they expire ITM).
• Practical use: Delta is also a rough proxy for the probability that an option expires ITM.

Gamma (Γ) — The Rate of Change of Delta

Gamma tells you how much delta will change for a $1 move in the underlying. It's the curvature of the option's value with respect to the stock price.

• Gamma is highest for ATM options near expiration.
• High gamma means your delta is unstable — small moves in the stock cause big changes in your directional exposure.
• Long gamma = you benefit from movement (long options). Short gamma = you benefit from stability (short options).
• Gamma is always positive for long options (both calls and puts).

Theta (Θ) — Time Decay

Theta measures how much value an option loses per day from the passage of time alone (all else equal). Time is the enemy of option buyers and the friend of option sellers.

• Theta is always negative for long options — they lose value every day.
• Theta accelerates as expiration approaches. The last month is brutal; the last week is devastating.
• ATM options have the highest theta. Deep ITM/OTM options have less time value to lose.
• Theta and gamma are inversely related: you can't have long gamma without paying theta, and you can't collect theta without being short gamma. This is a fundamental law of options.

Vega (ν) — Volatility Sensitivity

Vega measures how much the option price changes for a 1% change in implied volatility. This is perhaps the most important Greek at SIG, because SIG is fundamentally in the business of trading volatility.

• All options (long calls and long puts) have positive vega — they gain value when IV rises.
• Longer-dated options have higher vega — there's more time for volatility to matter.
• ATM options have the highest vega for a given expiration.
• When you "buy volatility," you're buying options (positive vega). When you "sell volatility," you're selling options (negative vega).

Rho (ρ) — Interest Rate Sensitivity

Rho measures sensitivity to interest rate changes. It's usually the least important Greek for short-dated options, but matters for LEAPS (long-term options) and in environments of rapid rate changes.

Second-Order Greeks

Professional traders also monitor higher-order Greeks:

Greek	What It Measures	Why It Matters
Vanna	∂Δ/∂σ or ∂ν/∂S — How delta changes with volatility	Critical for managing hedges when volatility shifts
Volga (Vomma)	∂²V/∂σ² — Rate of change of vega	Important for convexity of volatility exposure
Charm	∂Δ/∂t — How delta changes with time	Affects how you hedge overnight
Speed	∂Γ/∂S — Rate of change of gamma	Matters for very large positions in fast markets

Option Pricing Models

The Black-Scholes Model (1973)

The Black-Scholes model, developed by Fischer Black, Myron Scholes, and Robert Merton, is the foundational framework for pricing European options. The formula is:

Where:

• S₀ = current stock price
• K = strike price
• r = risk-free interest rate
• T = time to expiration (in years)
• σ = volatility of the underlying
• N(x) = cumulative standard normal distribution function

Key assumptions of Black-Scholes (and where it breaks down):

1. Constant volatility — In reality, volatility changes constantly. This is the biggest weakness.
2. Log-normal price distribution — Real markets have fat tails (extreme moves happen more often than the model predicts).
3. No transaction costs — Real trading has spreads, fees, and market impact.
4. Continuous hedging — You can't continuously rebalance; you hedge in discrete intervals.
5. European exercise only — American options can be exercised early.

The Binomial Model

The binomial model (Cox-Ross-Rubinstein, 1979) prices options by building a tree of possible price movements. At each step, the stock can move up or down by a fixed factor. It's more intuitive and handles American-style exercise naturally.

Put-Call Parity

For European options, there's a fixed mathematical relationship between call and put prices:

This is a no-arbitrage condition. If it's violated, there's free money on the table. Market makers like SIG constantly monitor put-call parity across thousands of options. Violations are tiny and fleeting — you need speed and technology to capture them.

Volatility: The Soul of Options

If there's one concept that defines options trading, it's volatility. At SIG, you're not really trading options — you're trading volatility.

Historical vs Implied Volatility

• Historical (realised) volatility: What the stock actually did — measured as the annualised standard deviation of past returns. It's backward-looking.
• Implied volatility (IV): What the market expects volatility to be — derived from current option prices using Black-Scholes in reverse. It's forward-looking.

The gap between implied and realised volatility is where market makers make or lose money. If you sell options priced at 30% IV but the stock only moves at 20% realised vol, you've pocketed the difference through delta hedging.

The Volatility Surface

In reality, implied volatility is not a single number — it varies by both strike price and expiration date, forming a three-dimensional surface.

• Volatility smile/skew (across strikes): OTM puts typically have higher IV than OTM calls for equities. This is called the skew. It reflects higher demand for downside protection and the empirical observation that markets crash more than they melt up.
• Term structure (across expirations): Normally, longer-dated options have higher IV. When short-term IV exceeds long-term IV, the term structure is inverted (backwardated) — usually signalling an expected near-term event like earnings or a Fed decision.

The VIX — The "Fear Index"

The VIX measures the market's expectation of 30-day volatility for the S&P 500, derived from SPX option prices. A VIX of 15 means the market expects the S&P to move about 15% annualised (or roughly ±1% per week). When the VIX spikes, fear is high; when it's low, complacency reigns.

Key Options Strategies

Strategy	Construction	What You're Betting On
Straddle	Long call + long put, same strike & expiry	Big move in either direction (long volatility)
Strangle	Long OTM call + long OTM put	Cheaper than straddle; needs a bigger move to profit
Vertical Spread	Buy + sell options at different strikes, same expiry	Directional with capped risk and reward
Calendar Spread	Buy + sell same strike, different expirations	Time decay or term structure changes
Iron Condor	Short strangle + long strangle wider	Stock stays in a range (short volatility)
Butterfly	Three strikes: buy 1 low, sell 2 middle, buy 1 high	Stock pins at middle strike
Collar	Long stock + long put + short call	Protection with capped upside
Ratio Spread	Buy n options, sell m options (n ≠ m)	Specific view on skew or vol surface

Market Making: The SIG Core Business

A market maker provides liquidity by continuously quoting two prices: a bid (the price they'll buy at) and an ask (the price they'll sell at). The difference is the bid-ask spread, and it's the market maker's compensation for taking on risk.

How Market Making Works

1. Quote: Post bid and ask prices for thousands of options simultaneously.
2. Fill: When someone trades against your quote, you accumulate a position.
3. Hedge: Use the underlying stock (or other options) to neutralize directional risk.
4. Manage: Continuously adjust hedges as prices move, Greeks change, and new information arrives.
5. Rinse & repeat: Thousands of times per day, across hundreds of underlyings, on 100+ exchanges.

The Market Maker's Enemies

• Adverse selection: Trading against someone who knows more than you. If an insider buys calls from you, you'll lose money on the hedge.
• Inventory risk: Accumulating too much of one position. If you're long 10,000 calls and the stock gaps down, you're in trouble.
• Model risk: Your pricing model is wrong. If IV should be 30% but your model says 25%, you're systematically underpricing the options you sell.
• Technology failures: If your system goes down, your quotes are stale, and you become a sitting target.

In modern markets, ideas are necessary but insufficient. The same mathematical insight, implemented with a 10-microsecond latency advantage, can mean the difference between profit and loss. SIG operates at the frontier of trading technology — not because speed alone makes money, but because speed enables the execution of strategies that are otherwise impossible.

What Is High-Frequency Trading?

HFT refers to trading strategies that operate at extreme speeds, holding positions for very short periods (microseconds to seconds) and executing large numbers of trades. But HFT is not a single strategy — it's a capability that enables many different strategies:

• Market making: Quoting bid-ask prices and profiting from the spread. Speed ensures your quotes are fresh and your hedges are timely.
• Statistical arbitrage: Exploiting tiny mispricings between correlated securities. These disappear in milliseconds.
• Latency arbitrage: Being faster than other participants to react to new information (controversial and increasingly difficult).
• Event-driven: Reacting to economic data releases, earnings announcements, or order flow patterns before others.

The Infrastructure Stack

Co-location

Co-location means placing your servers in the same data centre as the exchange's matching engine. When the speed of light through fibre optic cable is approximately 200,000 km/s (about 5 microseconds per kilometre), every metre of cable matters.

• Major exchanges (NYSE, NASDAQ, CME) offer co-location services in their data centres.
• Firms pay significant premiums for the closest possible rack positions.
• Cross-connects (direct cables between your server and the exchange) are measured in metres.
• SIG co-locates at all major global exchanges where they trade.

Network Infrastructure

• Microwave/millimetre-wave links: Between major exchanges (e.g., Chicago CME to New Jersey NYSE), microwave towers transmit data faster than fibre because microwaves travel closer to the speed of light in vacuum (~300,000 km/s vs ~200,000 km/s in fibre). The New York–Chicago route is approximately 1,000 km; microwave saves roughly 2-3 milliseconds over fibre.
• Kernel bypass networking: Technologies like DPDK, Solarflare OpenOnload, and Mellanox VMA allow network packets to reach your application directly, bypassing the operating system kernel's network stack. This eliminates microseconds of overhead per packet.
• Multicast market data: Exchanges broadcast market data via multicast — one transmission reaches all subscribers simultaneously, ensuring equal delivery times.

FPGAs — Field Programmable Gate Arrays

FPGAs are semiconductor chips that can be programmed at the hardware level to perform specific computations. Unlike CPUs (which execute instructions sequentially) or GPUs (which excel at parallel floating-point work), FPGAs can be configured for deterministic, parallel processing of specific tasks.

• Latency: FPGAs can process market data and generate orders in as little as hundreds of nanoseconds.
• Determinism: No operating system, no garbage collection, no cache misses. The latency is consistent.
• Applications: Market data parsing, risk checks, simple order generation, packet processing.
• Limitation: Harder to program than software. Complex strategies still run on CPUs with FPGA assistance for the most latency-critical components.

Hardware Timestamping

Precise timing is fundamental. Network cards with hardware timestamping can capture packet arrival times with nanosecond precision, enabling firms to measure and optimise latency with surgical accuracy. GPS-synchronised clocks ensure all systems share a common time reference.

Market Microstructure: How Markets Really Work

Market microstructure is the academic field that studies how trading mechanisms affect price formation, liquidity, and transaction costs. For an HFT trader, microstructure theory is the playbook.

The Limit Order Book

The order book is the central data structure of electronic markets. It contains all resting buy orders (bids) and sell orders (asks) at every price level, along with their sizes.

• Best bid / best ask: The highest bid and lowest ask. The spread between them is the bid-ask spread.
• Depth: The quantity available at each price level. Thin books mean large orders will move the price.
• Price-time priority: Most exchanges match orders first by price (best price wins), then by time (first in queue wins at the same price).

Order Types

Order Type	Behaviour	Who Uses It
Market order	Executes immediately at best available price	Aggressive traders who want certainty of execution
Limit order	Rests in the book until the price is reached or it's cancelled	Passive traders / market makers
IOC (Immediate or Cancel)	Fill what's available now, cancel the rest	Aggressive electronic traders
FOK (Fill or Kill)	Fill the entire order or cancel it all	Large block trades
Iceberg	Shows only a portion of the total size	Large institutional orders hiding their size

The Matching Engine

The matching engine is the heart of an exchange. It receives incoming orders, matches buyers with sellers based on priority rules, and generates execution reports. Modern matching engines operate with latencies of single-digit microseconds. CME's Globex engine processes millions of messages per second.

Maker-Taker Fees

Many exchanges use a maker-taker fee model: they pay traders who add liquidity (makers — those posting limit orders) and charge traders who take liquidity (takers — those sending market orders). This incentivises market making and tighter spreads.

The Latency Arms Race

Measuring Latency

Metric	What It Measures	Why It Matters
Tick-to-trade	Time from receiving market data to having an order on the wire	The single most important speed metric
Average latency	Mean response time	Useful but can hide problems
Tail latency (p99)	The slowest 1% of responses	Often more important than average — your worst moments define your losses
Jitter	Variance in latency	Inconsistency makes strategies unreliable

Physics of Speed

• Speed of light in vacuum: ~300,000 km/s (299,792 to be precise)
• Speed of light in fibre: ~200,000 km/s (refractive index ≈ 1.5)
• Speed of microwave in air: ~299,000 km/s (very close to vacuum)
• New York to Chicago (fibre): ~6.5 ms round trip
• New York to Chicago (microwave): ~4.3 ms round trip
• London to Frankfurt (fibre): ~4.2 ms round trip

Software Engineering for Trading

The Hot-Cold Architecture

Trading systems separate the hot path (latency-critical: market data processing, signal generation, order routing) from the cold path (everything else: logging, risk reporting, P&L calculation, analytics).

• The hot path is written in C++ (or uses FPGAs) with extreme attention to memory allocation, cache utilisation, and branch prediction.
• The cold path can use Python, Java, or any language — performance doesn't matter here.
• The two communicate via shared memory, ring buffers, or messaging systems.

Key Engineering Principles

• Zero allocation on the hot path: Memory allocation triggers garbage collection (in managed languages) or system calls. Pre-allocate everything.
• Lock-free data structures: Mutexes cause threads to wait (blocking). Lock-free structures use atomic operations so threads never block, giving consistent latency.
• Cache-friendly data layout: CPU cache misses cost ~100ns each. Organise data so that frequently accessed fields are adjacent in memory.
• Branch prediction awareness: Modern CPUs predict which branch of an if-statement will be taken. Mispredictions cost ~15-20 cycles.
• Busy-wait over sleep: Sleeping a thread yields CPU to the scheduler; waking up adds microseconds. Hot path threads busy-wait (spin) on dedicated CPU cores.

Smart Order Routing (SOR)

When a stock trades on multiple exchanges (as most do), SOR algorithms determine the best venue for each order. They consider:

• Price and available liquidity at each venue
• Exchange fees (maker-taker, payment for order flow)
• Latency to each venue
• Fill probability and information leakage risk
• Regulatory requirements (best execution obligations)

SIG doesn't just use game theory as a metaphor — they live it. New traders learn poker before they learn to trade. The firm was founded by poker players. Every trading decision is a decision under uncertainty with incomplete information, and game theory provides the rigorous framework for making those decisions optimally.

Foundations of Game Theory

What Is Game Theory?

Game theory is the mathematical study of strategic interaction — situations where your optimal decision depends on what others are doing. Developed by John von Neumann and Oskar Morgenstern (1944) and revolutionised by John Nash (1950), it provides frameworks for understanding competitive and cooperative behaviour.

Key Concepts

Nash Equilibrium: A set of strategies where no player can improve their outcome by unilaterally changing their own strategy. It's a stable state — everyone is doing the best they can given what everyone else is doing.

• In the classic Prisoner's Dilemma, mutual defection is the Nash Equilibrium (even though mutual cooperation would be better for both).
• Markets often settle into Nash Equilibria — for example, bid-ask spreads tend to reach levels where no single market maker can profitably deviate.

Dominant Strategy: A strategy that's optimal regardless of what your opponents do. If you have one, always play it. In trading, truly dominant strategies are rare — most situations require conditional reasoning.

Zero-Sum vs Positive-Sum Games:

• Zero-sum: One player's gain is another's loss. Short-term trading and poker are approximately zero-sum.
• Positive-sum: Total value can grow. Long-term investing is positive-sum (companies create value). Market making is arguably positive-sum because it provides liquidity that benefits all participants.

Complete vs Incomplete Information:

• Complete information: Everyone knows the rules, payoffs, and others' options (e.g., chess).
• Incomplete information: Hidden information exists (e.g., poker — you can't see others' cards). Trading is almost always incomplete information.

Mixed Strategies: Instead of always choosing one action, you randomise between actions with specific probabilities. Nash showed that every finite game has at least one equilibrium, possibly in mixed strategies. In trading, this means sometimes the optimal response isn't deterministic — it's probabilistic.

The Prisoner's Dilemma and Repeated Games

In the classic single-round Prisoner's Dilemma, defection dominates. But in repeated games (where you play many rounds against the same opponent), cooperation can emerge. Robert Axelrod's famous tournament (1984) showed that tit-for-tat — start by cooperating, then mirror your opponent's last move — outperformed all other strategies.

In trading, many interactions are repeated. If you consistently trade fairly with a counterparty, they'll continue providing good prices. If you exploit someone once, they'll widen their spreads against you or stop trading with you entirely. Reputation matters.

Auction Theory

Markets are essentially continuous auctions. Key concepts:

• Winner's curse: If you win an auction (or if you're the one filling a large order), you bid the highest — which should make you wonder if you overvalued what you bought. Correcting for winner's curse is essential in market making.
• Revenue equivalence: Under certain conditions, different auction formats generate the same expected revenue. But in practice, the format matters enormously for trading strategy.
• Information revelation: Your actions reveal information about your beliefs. Market makers must balance quoting aggressively (to win trades) with quoting conservatively (to avoid revealing too much information).

Expected Value: The Master Concept

Expected value (EV) is the single most important concept in trading, gambling, and decision-making under uncertainty. It is the probability-weighted average of all possible outcomes.

EV in Action

Example 1: A coin flip pays $20 on heads, costs $10 on tails.

This is a positive EV bet. Take it every time. Over thousands of flips, you'll average $5 per flip.

Example 2: You roll a fair 6-sided die. You win the face value in dollars. How much would you pay to play?

Pay anything less than $3.50 and you have a positive EV game.

Example 3: An option trade has a 30% chance of making $100,000 and a 70% chance of losing $20,000.

Despite losing 70% of the time, this is a great trade. This is exactly how SIG thinks. They don't ask "will I win this trade?" They ask "is the expected value positive?"

Poker and Trading: The Deep Connection

SIG teaches poker to new traders not as a perk, but as curriculum. The firm is home to World Series of Poker Bracelet winners Bill Chen and Jerrod Ankenman, who co-authored The Mathematics of Poker. Here's why the connection runs so deep:

Concept	In Poker	In Trading
Expected value	Calculate pot odds vs hand odds for every decision	Calculate EV of every trade vs its risk
Incomplete information	You can't see opponents' cards	You don't know others' positions, intentions, or information
Position sizing	Bet sizing relative to pot and stack	Trade sizing relative to portfolio and risk budget
Pot odds	Reward/risk ratio of calling a bet	Risk/reward ratio of entering a trade
Reading opponents	Betting patterns reveal hand strength	Order flow patterns reveal informed trading
Tilt	Emotional play after bad beats	Revenge trading after losses
Bankroll management	Don't risk going bust	Don't risk blowing up
Variance tolerance	Losing sessions don't mean losing strategy	Losing days don't mean losing edge
Range-based thinking	Opponent has a range of possible hands	Stock has a range of possible outcomes
Game selection	Play tables where you have an edge	Trade markets where you have an edge

Pot Odds: A Trading Metaphor

In poker, pot odds compare the cost of calling to the potential reward:

If the pot is $100 and you must call $20, your pot odds are 20/120 = 16.7%. You need at least a 16.7% chance of winning to make the call profitable.

In trading: if you risk $10,000 on a trade that could make $40,000, you need at least a 20% win probability (10,000/50,000) for positive EV. This is the same calculation.

Bayesian Thinking: Updating Beliefs

Bayesian inference is the mathematical framework for updating your beliefs as new evidence arrives. It's central to how SIG traders think.

In words: the probability of A given B equals the probability of B given A, times the prior probability of A, divided by the probability of B.

Trading Application

Suppose you believe a stock has a 60% chance of going up tomorrow (your prior). Then you see unusual options activity — heavy call buying. This is new evidence. How do you update?

• If heavy call buying is more likely in a "stock going up" world, your posterior probability should increase above 60%.
• If heavy call buying is common regardless of outcome (maybe it's just retail activity), it shouldn't change your belief much.
• The key is the likelihood ratio: how much more likely is this evidence under one hypothesis vs the other?

Good traders are Bayesian updaters — they constantly revise their views based on new data. Bad traders are either too stubborn (ignoring new evidence) or too reactive (overweighting every new piece of information).

Risk Management: Sizing Your Bets

The Kelly Criterion

Kelly (1956) solved the problem of optimal bet sizing: how much of your bankroll should you risk on each positive EV bet to maximise long-term growth?

Where f* = fraction of bankroll to bet, b = net odds (win amount / loss amount), p = probability of winning, q = 1 - p.

Example: A bet wins $200 with 60% probability and loses $100 with 40% probability.

Kelly says bet 40% of your bankroll. In practice, most professionals use fractional Kelly (1/2 Kelly or 1/4 Kelly) to reduce volatility and protect against estimation errors.

Risk of Ruin

Even with a positive EV strategy, bad position sizing can lead to ruin. Risk of ruin is the probability of losing everything. Key insight: the risk of ruin is non-zero even with a positive edge if your bets are too large.

Why Many Small Bets Beat Few Large Bets

With a positive edge, the law of large numbers works in your favour over many trials. The more bets you make, the more your actual results converge to the expected value. This is why diversification works, why market makers take thousands of small positions instead of a few big ones, and why professional poker players play thousands of hands.

SIG's stated belief: "Smart people working together always come up with a better solution than the lone genius." In a firm where traders, quantitative researchers, and software engineers must collaborate in real-time on systems that trade billions of dollars, the ability to work across boundaries isn't a soft skill — it's a survival skill.

How Trading Firms Are Structured

The Three Pillars

Most quantitative trading firms organise around three core functions:

1. Trading: Making markets, managing positions, and making real-time decisions. Traders understand market dynamics, customer flow, and have the instincts forged by experience.
2. Quantitative Research: Building models for pricing, risk, signal generation, and portfolio optimisation. Quants bring mathematical rigour and statistical methods.
3. Technology: Building and maintaining the infrastructure — market data systems, execution platforms, risk engines, and the low-latency stack. Engineers make everything work at scale and speed.

Why Silos Are Dangerous

• Traders without quants make decisions on gut feeling, missing systematic patterns.
• Quants without traders build elegant models that don't work in real markets (ignoring liquidity, market impact, and the messy reality of execution).
• Engineers without domain context optimise the wrong things or build systems that don't match how traders actually work.
• The best solutions emerge at the intersections — a trader noticing a pattern, a quant formalising it, and an engineer implementing it fast enough to profit.

Communication Across Domains

The Translation Problem

Each function speaks its own language. Traders talk about "flow," "skew," and "axe." Quants talk about "alpha," "Sharpe ratio," and "cross-validation." Engineers talk about "throughput," "p99," and "cache misses." The most valuable team members are bilingual — they can translate between domains.

Principles of Effective Cross-Team Communication

1. Lead with the problem, not the solution. Instead of "we need a faster database," say "our lookup latency is causing stale quotes in fast markets, costing us $X per day."
2. Quantify everything. "This is slow" is useless. "Our p99 latency is 200μs but needs to be under 50μs to compete at these exchanges" is actionable.
3. Understand constraints. Engineers can't make physics go faster. Traders can't predict the future. Quants can't model what they can't measure. Respect each domain's genuine limitations.
4. Shared vocabulary. Invest time in learning enough of other domains to have meaningful conversations. You don't need to be an expert, but you need to be conversant.
5. Intellectual honesty. If you're wrong, admit it quickly. If you don't know something, say so. Pretending creates compounding errors.

Decision Rights and Escalation

Clear decision rights prevent paralysis. In trading firms:

• Real-time trading decisions: The trader on the desk has authority. Speed matters more than consensus.
• Model changes: Require quant + trader agreement. A model that a trader doesn't trust won't be used.
• Infrastructure changes: Engineering leads, but must communicate impact and timeline to trading.
• Strategy changes: Collaborative. The best ideas emerge from all three functions debating together.

Productive Conflict and Debate

At SIG, disagreement is not just tolerated — it's expected. If everyone agrees, someone isn't thinking hard enough.

The "Red Team" Mentality

Before implementing a new strategy or significant change, teams actively try to break their own ideas. This involves:

• Assigning someone to argue the opposite position (devil's advocate)
• Stress-testing assumptions: "What if volatility doubles? What if liquidity disappears? What if our model is wrong by 20%?"
• Post-mortems after failures — not to assign blame, but to understand what went wrong and improve the process

Psychological Safety

For productive conflict to work, people need to feel safe challenging ideas — including ideas from senior people. SIG's culture of "questioning assumptions" only works if junior team members genuinely believe they won't be punished for pushing back. This is a deliberate cultural investment.

Cross-Functional Project Success Patterns

Pattern	What It Looks Like
Embedded team members	An engineer sits with the trading desk. A quant spends time in the tech war room.
Shared metrics	Everyone is measured on the same P&L, not separate function-level KPIs.
Joint retrospectives	All three functions review what happened, not just their own piece.
Rotational programs	SIG rotates new hires through different desks and functions.
Social integration	Games nights, sports teams, shared meals. Relationships built outside of pressure situations transfer to pressure situations.

Susquehanna International Group (SIG) is one of the largest privately held financial firms in the world. Understanding the firm's culture, values, and working environment is just as important as understanding the technical material — perhaps more so, because culture fit is what separates a good candidate from a great one.

Origins: From Poker Table to Global Trading Floor

In the early 1980s, a group of friends from college began trading independently on the floor of the Philadelphia Stock Exchange. What set them apart wasn't finance degrees or Wall Street pedigrees — it was their quantitative skills and poker experience. They understood probability, expected value, and risk management intuitively from thousands of hours at the poker table.

Realising they could be more successful working collaboratively, the founders joined together in 1987 to start what is now Susquehanna International Group. The firm's name comes from the Susquehanna River in Pennsylvania, near their headquarters in Bala Cynwyd (a suburb of Philadelphia).

Today, SIG:

• Trades on more than 100 global exchanges
• Is one of the world's largest options market makers
• Has major offices in Philadelphia, Dublin (Ireland), Sydney, and other locations
• Has significant investments including a major stake in ByteDance (TikTok's parent company)
• Uses proprietary computer models to make trading decisions
• Maintains its founding culture of games, collaboration, and quantitative rigour

The Gaming Culture: More Than Fun

Games aren't a corporate perk at SIG — they're infrastructure. The firm uses games to develop and assess the skills that matter most in trading.

Poker

SIG's signature game. New traders go through poker training as part of onboarding. The firm is home to WSOP Bracelet winners Bill Chen and Jerrod Ankenman, co-authors of The Mathematics of Poker. Poker teaches:

• Decision making under uncertainty (exactly like trading)
• Expected value calculation in real time
• Emotional discipline (managing tilt)
• Reading opponents (understanding order flow)
• Bankroll management (position sizing)

Chess

Chess teaches decision tree pruning under time pressure. You can't calculate every possible branch, so you must identify the most important features of a position and form a plan quickly. This maps directly to trading, where you can't process all available information — you must focus on what matters most.

Magic: The Gathering

MTG requires decision making with imperfect information, observation of opponents, and strategic resource management. SIG takes it seriously enough that their head of trading in Sydney is a professional Magic player.

Wizard (Card Game)

Wizard is a trick-taking game with a betting element. It's "easy to learn and hard to master" — SIG uses it to introduce candidates to gaming during the interview process.

Strategic Board Games

Catan, Dominion, Backgammon, Hanabi, Power Grid, Love Letter, and The Resistance: Avalon are all popular at SIG. Each develops specific skills: negotiation (Catan), probability (Backgammon), cooperation with hidden information (Hanabi), and deception detection (Avalon).

Esports

SIG values action games that require reactions to small changes in a fast-paced environment — a direct parallel to monitoring market data feeds and reacting to price changes.

The SIG Interview Process

Stage 1: Online Assessment

The SIG Quantitative Evaluation (20 minutes for internships, 60 minutes for full-time). Tests logic, probability, expected value, and puzzle-solving. Challenges include:

• Solving linear systems of equations
• Calculating expected values
• Complex puzzles (grid path-finding, combinatorics)
• Brain teasers with quantitative elements

Stage 2: Phone Interview

Covers your background, motivation ("Why SIG?"), market awareness, and math problems (probability, conditional probability, combinatorial analysis). Key points:

• Have a genuine, specific reason for wanting SIG (mention their game theory approach, culture, specific aspects)
• Follow financial news (Bloomberg, FT, etc.)
• Show your thinking process, not just answers

Stage 3: Technical Video Interview

A 60-minute session with a senior trader. Primarily probability-based math questions, plus CV discussion. Expect questions that build on each other, getting progressively harder.

Stage 4: Super Day (Final)

Multiple interviews covering:

• Expected value games: With cards, dice, or coins. You'll make trading decisions in a simulated market-making environment. They assess your math, composure, and communication.
• Card/dice games: Real games played with interviewers. Wizard is commonly used. They watch how you think, bet, and handle uncertainty.
• Behavioural interviews: Teamwork, communication, cultural fit.

SIG's Core Values

Value	What SIG Says	What It Means in Practice
Good Decision Making	"A rigorous and analytical approach... requires reflection and practice."	Data-driven. Challenge assumptions. Post-mortem failures. Process > outcome.
Teamwork	"Smart people working together always beat the lone genius."	Collaboration isn't optional. Best ideas emerge from diverse teams debating.
Belonging	"A culture of collaboration that encourages every employee to feel welcome."	Psychological safety. Diversity of thought. Ask questions freely.
Growth Mindset	"Constantly learning and challenging ourselves to grow."	Nobody is "done" developing. Learn from mistakes. Seek feedback.
Expertise	"We value the expertise of the people we hire."	Deep domain knowledge matters. Each person has meaningful impact.
Integrity	"Dependability and honesty are the foundation of our relationships."	Don't hide losses. Don't fudge numbers. Trust is everything.

What Glassdoor & Employee Reviews Say

• Ratings: ~75% of employees recommend working at SIG. 3.8/5 for work-life balance, 3.9/5 for culture.
• Positives frequently mentioned: Collaborative culture, game nights, intellectual stimulation, training programs, competitive compensation, smart colleagues.
• Negatives sometimes mentioned: Demanding hours during market events, steep learning curve, can feel intense.
• The consistent theme: SIG attracts people who genuinely enjoy solving hard problems and competing — not against each other, but against the market.

Here's the thing that separates a good interview candidate from a great one: understanding that these five subjects aren't separate — they're one thing viewed from five angles.

A SIG trader sits at a desk, watching options prices on a screen (Chapter 1), delivered by an infrastructure stack that processes millions of data points per second (Chapter 2). They make buy and sell decisions based on expected value calculations and game-theoretic reasoning about who's on the other side of the trade (Chapter 3). They collaborate with engineers and quants to improve the models and systems that give them an edge (Chapter 4). And they do all of this within a culture that treats games as training, debate as essential, and intellectual honesty as sacred (Chapter 5).

Every one of these threads reinforces the others:

• Options theory gives you the "what" — what to trade, how to price it, what risks you're taking.
• Technology gives you the "how" — the speed, reliability, and scale to execute.
• Game theory gives you the "why" — the decision framework for when to act and how much to risk.
• Cross-team work gives you the "who" — the people you need and how to work with them.
• SIG's culture gives you the "where" — the environment that makes all of it possible.

Walk into that interview knowing this, and you won't just answer questions. You'll speak their language.