Susquehanna Poker Showdown Recap: When the prize is not the prize

Approximately 1,400 people entered the Susquehanna Poker Showdown process. The competition began with an online estimation qualifier, from which around 100 people advanced to a live poker tournament. The final two tables, representing the top 18 finishers, earned an interview opportunity with Susquehanna.

The advertised prize pool was $12,500. I was happy to win whatever part of it came my way, but I had not spent my evenings after work studying satellite strategy for an expected share of $125.

The interview was the opportunity that mattered.

Susquehanna is one of the few firms where my background in poker is directly relevant to the way the firm describes trading, probability and decision-making. I had spent years making real-money decisions under uncertainty before moving into data science, where many of the same instincts became formal models, code and statistical analysis. The Showdown sat across both parts of my life and offered a rare recruitment path where that background could be demonstrated through an actual competition rather than compressed into a line on a CV.

That made the event personally important to me. It also meant I was never going to approach it casually.

Poker would retain its usual right to ignore my plans and eliminate me anyway. I could still remove as many avoidable ways of losing as possible. If the opportunity was going to be taken from me, I wanted that to happen while I was performing at my best.

The Competition Began Before the Cards

The first stage was an online estimation qualifier based on Fermi-style questions. These ask for quantities that are difficult or impossible to know directly, requiring the problem to be broken into smaller components that can be estimated and combined.

A good response depends on choosing sensible assumptions, identifying which quantities drive the result and keeping the arithmetic under control. The answer may be uncertain, but the reasoning should still be defensible.

That format appealed to me. It combines modelling, probabilistic judgement and the willingness to make a decision without complete information. It also punishes the comforting habit of refusing to answer until every input is known, which would be an unfortunate characteristic in both poker and trading.

When I first saw the qualifier, I studied how Fermi questions were normally approached and began practising interval estimates. My assumption was that a strong response would provide a plausible range reflecting the uncertainty in the inputs. When further research suggested that the Showdown qualifier was likely to require a single point estimate, I changed the way I trained and practised committing to one answer.

That adjustment became the pattern for the rest of the competition. I wanted to understand the game that was actually being played, rather than prepare for the version I happened to expect.

The estimation challenge selected roughly 100 live finalists from the original field of approximately 1,400. Once the tournament invitation arrived, the problem changed completely. Estimation had earned me a seat in the room; poker would determine whether I reached the interview threshold.

Preparing for the Game That Mattered

My preparation focused on tournament structure, the likely field and the practical conditions under which I would have to make decisions for most of a day.

I began studying tournament poker every evening after work and throughout the weekends. I watched footage from the American Susquehanna Poker Showdown, searched for anything previous competitors had written and tried to understand how an intelligent but mixed-experience field would respond to a fast structure and an unusually valuable bubble.

My own poker background was mainly in cash games, especially Limit Hold’em. Cash experience gave me comfort with variance, long sessions and folding without becoming impatient, but those strengths did not automatically transfer to a rapidly escalating tournament. I needed to rebuild the parts that were specific to stack depth, fold equity, satellites and bubble pressure.

The practical preparation had a different purpose. A week before the tournament, I played an eight-hour cash session at the casino to reacclimatise myself to the physical rhythm of live poker. I wanted handling chips, watching the action and making uninterrupted decisions for hours to feel ordinary again.

The aim was never to pretend I would feel no nerves. I wanted the room and the mechanics of live play to stop being novel before the tournament began.

I also took the day before the event off work and travelled to the building. I checked the route, the travel time and the entrance so that none of those details required thought on the morning itself. Getting lost before a poker tournament would have been a memorable story, but not one that improved my expected value.

Food received the same treatment. I planned what I would eat the night before and on the morning of the event with cognitive endurance in mind, then brought Up&Go, Red Bull and biltong in case suitable food or coffee was unavailable. I was not going to allow a missing sandwich to become a strategic variable.

My wife has periodically reminded me that not everything is a competition. She is probably correct in a broad philosophical sense. The difficulty was that this one clearly was.

Running Deep While Card Dead

The preparation mattered because the deck offered almost no assistance during the tournament.

The best starting hand I received all day was AQ. There was no run of premium pairs, no effortless double-up and no period where the cards reduced the problem to waiting for opponents to pay me.

I still reached the final two tables and finished inside the top 18.

I mention the card distribution because the strategy had to function through the branch where the cards remained poor and the blinds continued rising. Preparing only for the version of the tournament where I found premiums, doubled early and coasted towards the threshold would have achieved very little. That version largely takes care of itself.

The difficult version requires the player to preserve fold equity, identify uncontested pots and remain capable of applying pressure while the stack is steadily being taxed.

That was the version I received.

The Prize Was Not the Prize

The first strategic question was:

What game am I actually playing?

The advertised prize pool was $12,500. With roughly 100 live finalists, the average prize value per entrant was approximately $125. Even after assigning myself a meaningful edge over the field, I estimated my cash-prize expected value at around $250.

The event was likely to occupy most of the day, so the prize component was worth roughly $30 per hour by my estimate. I have extensive recorded live 5/5 cash-game results with a long-term hourly rate above that figure.

The money remained welcome, but it could not explain the amount of work I was putting into the event. The interview threshold could.

Reaching the final two tables would place me inside the top 18 and secure the opportunity to speak with Susquehanna. Once I understood that objective clearly, the first phase of the tournament looked much more like a satellite than a conventional tournament.

In a satellite, the goal is to survive until a seat or qualification place has been secured. Finishing with every chip provides no additional benefit over finishing with enough chips to qualify. The value of accumulation therefore depends on how much it improves the probability of reaching the threshold.

Before the interview bubble, chip accumulation mattered because it supported survival, preserved leverage and reduced the risk of elimination. Survival itself carried a value that ordinary chip EV did not capture.

After the threshold, the same chips would be valued differently. The interview would already be secured, and the remaining tournament would have a top-heavy payout structure where accumulation once again became the priority.

The event therefore contained two strategic phases:

• Before the interview threshold: play the incentives of a satellite.
• After the interview threshold: play a top-heavy poker tournament.

The ranges and appetite for variance needed to change when the objective changed.

Studying the First Game

Once I understood that the early tournament functioned as a satellite, I concentrated heavily on Dara O’Kearney’s satellite material. Satellite strategy treats survival and chip accumulation differently because the final objective is qualification rather than ownership of every chip.

This changes the risk attached to all-in decisions.

When I shove first, I can win the pot uncontested. The opponent has to decide whether the strength of their hand justifies risking chips or tournament life. When I call an all-in, the possibility of winning through a fold has disappeared. I realise my equity and accept the full variance of the confrontation.

That asymmetry became central to my plan. I wanted to create difficult decisions for other players while remaining extremely selective about volunteering for the decisions they created for me.

I also revisited The Mathematics of Poker, particularly the material on short-stack play and the compression of the decision tree as stacks become shallower. At around ten big blinds, much of the postflop game disappears, and decisions increasingly revolve around open shoves, re-shoves, calls and folds.

Upswing Lab and other tournament material helped translate those ideas into practical ranges and stack-depth adjustments. I was not going to become a tournament specialist in several weeks, but I could avoid bringing unchanged cash-game instincts into a fast tournament and hoping they would remain appropriate.

I also estimated the stack likely to be required to reach the final two tables. My estimate was approximately 3.5 starting stacks. With a 5,000-chip starting stack, I expected roughly 17,500 chips to be sufficient, subject to the elimination rate and the field’s willingness to protect tournament life.

I reached the final two tables with approximately 20,000 chips, or four starting stacks. The estimate was close, and the broad strategy had produced the stack it was intended to produce.

Profiling the Field

I used the American Showdown footage to build an initial population model.

The contestants were clearly intelligent, numerate and accustomed to thinking quickly. Some appeared to have limited poker experience, although that did not make them harmless. Their mistakes would simply have a different shape from those made by regular poker players.

Raw intelligence was unlikely to be my edge in that room. It was the entry requirement.

My edge had to come from understanding the specialised decision environment: tournament incentives, stack leverage, fold equity, satellite risk, live pressure and the way those factors interact near a meaningful bubble.

That assessment held up during the event. I was sitting with some of the smartest students in the country. Many were quick to absorb information and adjust, even when they lacked extensive poker experience.

The useful starting model was a field of highly intelligent people playing a specialised game in which poker experience and comfort under direct pressure still mattered.

I expected three broad player types.

Seat-preservers would understand the value of the interview threshold and attempt to fold their way towards it. Their mistake would be treating inaction as safety even while increasing blinds steadily removed their leverage. First-in pressure would become especially valuable against them.

Splashy or performative players would want to demonstrate confidence, play conspicuous pots or gamble in situations where activity felt more impressive than restraint. I had no reason to fight ego with ego. I could avoid marginal confrontations and allow them to provide action when my range justified it.

Prepared tournament-aware players would understand stack pressure and bubble incentives. They could not be treated as automatic chips, and lazy aggression against them would be expensive.

Several students had clearly prepared with the same seriousness I had. That was easy to recognise, and I respected it. They had understood that preparation formed part of the competition even though it did not appear in the blind structure.

Others had arrived with less preparation. The tournament was likely to explain the omission to them more efficiently than I could.

These categories were only starting priors. Every open, fold, call, hesitation, showdown and stack-protection decision provided new evidence. The job was to begin with a useful model and update it quickly.

Turning Ranges into a Policy

Before the tournament, I prepared a short-stack chart containing tight, medium and loose range buckets for open shoves, re-shoves and call-offs.

The chart provided a library of possible ranges. It could not select the right one for a particular hand because that decision depended on the state of the tournament.

Before looking at my cards, I considered the number of players remaining, my distance from the interview bubble, my stack depth, the stacks behind me, the likelihood that the blinds were protecting their seats, the amount of fold equity I retained and the damage a loss would do to my qualification probability.

I also had to identify which of the tournament’s two games I was currently playing. Before the threshold, the state generally selected tighter call-off ranges and placed a premium on first-in pressure. After qualification, the same range library could support a more aggressive accumulation policy.

Once the tournament state selected the appropriate range, the cards determined whether the current hand fell inside it.

The purpose of this preparation was to avoid looking down at a marginal hand and beginning an emotional negotiation with myself while the clock was running. I wanted the important strategic choice to have been considered before the cards supplied the emotional stimulus.

Tournament state came first, followed by the cards.

Card Dead Was Part of the Distribution

Having played hundreds of thousands of poker hands, I understand that card-dead stretches belong to the distribution. Twenty poor hands remain noise. They feel more significant in a fast tournament because the blinds are increasing while the player waits and every orbit removes future flexibility.

Being card dead gradually changes the tournament state. The stack declines, opportunities to win at showdown become rarer, and the player develops a tight image whether or not that was the original intention.

A tight image can create fold equity while the stack remains large enough to threaten other players. Once the stack has lost that leverage, the image becomes little more than a historical record of hands folded.

I could not allow the structure to remove every useful weapon while I waited for cards that might never arrive.

Mollitia Exitium Est

Whenever my stack approached the danger zone, I kept returning to one phrase:

Mollitia exitium est.

Comfort leads to ruin.

The comfortable decision was usually to wait for a stronger hand, a cleaner situation or some overdue cooperation from the deck. Folding felt disciplined, while elimination after forcing the action would have felt foolish.

Cash-game poker had made me extremely comfortable with folding. I can fold for hours without becoming impatient, which is valuable when the blinds remain fixed and another hand will arrive under the same conditions.

A fast tournament is less accommodating. The blinds attack the stack whether the player acts or waits, and every fold can reduce the leverage available in later hands.

One of my normal strengths could therefore become a weakness. I was capable of remaining calm while surrendering fold equity, stack pressure and future decision rights. Comfort could present itself as discipline until the stack became too small to make anyone fold.

When the tournament state required aggression, I followed the policy despite the discomfort attached to it. The strategy could be audited afterwards, when fear, fatigue and the immediate possibility of elimination were no longer influencing the answer. Abandoning the policy whenever execution felt unpleasant would have reduced it to a description of my mood.

Keeping the Stack Alive

Satellite strategy does not consist of sitting motionless while hoping that other people eliminate one another.

Earlier in the tournament, I won several postflop pots with no meaningful hand. They were unremarkable in isolation, but they kept the stack above the zones where fold equity would begin to collapse. I could not continue waiting for premiums while the blinds steadily converted my stack into something opponents could call without concern.

Those pots came from position, board texture, opponent type and the story created by the preceding action. I remained selective, but I applied pressure when the situation justified it.

One of my main rules was to avoid falling below ten big blinds whenever I could reasonably prevent it. At around 10bb, an open shove still carries enough force to make an opponent fold a hand with meaningful equity. The stack can win the blinds without showdown and impose a real cost on a caller.

At 5bb, opponents can call much more freely. Fold equity declines sharply, the decision tree contracts and the player becomes increasingly dependent on the cards.

A stack therefore represents more than its chip count. It contains future decision rights.

When my stack moved towards the 10bb threshold with around 40 players remaining, I deliberately widened my stealing frequency. The cash-game part of my brain was comfortable continuing to fold, but the tournament state made that increasingly expensive. I remained far enough from the interview bubble that waiting could destroy my leverage, while my stack was still large enough to make potential callers uncomfortable.

At the same time, my call-off range remained extremely tight. I folded hands that I might have called in a conventional tournament because the satellite risk premium was too large. First-in pressure could win without showdown, while calling an all-in accepted the full risk with no fold equity remaining.

I wanted to create the difficult decisions wherever possible rather than accept them unnecessarily.

The Interview Bubble

As the field approached the final two tables, I tightened substantially once my position appeared secure.

When 20 players remained and my stack was comfortably within the likely qualification range, playing a significant pot carried an enormous opportunity cost. I would have folded pocket aces to a shove in the relevant state.

Doubling my stack would still have earned the same interview, while losing would have thrown away an opportunity that was close to secured. The additional chips had little value before the threshold, whereas elimination retained its full cost.

That was the satellite logic in its clearest form.

I reached the final two tables with approximately 20,000 chips from a starting stack of 5,000. Four starting stacks placed me above the 3.5-stack estimate I had made before the event.

The first phase of the strategy had worked, and the interview was secured.

The Second Game

After the final two tables were reached, survival lost much of its earlier premium. I had a below-average stack in a top-heavy payout structure, which made chip accumulation substantially more valuable.

I widened my ranges and became more willing to accept volatile situations that offered a reasonable chance of increasing the stack. The underlying range library remained the same, while the tournament state now selected a more aggressive policy from it.

Continuing to protect the interview after it had already been secured would have meant solving a problem that no longer existed.

The Final Hand

My final hand was A6 suited into pocket kings.

Given the same position, stack and tournament state, I would make the open shove every time. The interview threshold had already been secured, my stack was below average and accumulation had become the objective.

A suited ace blocks some ace-heavy calling combinations, retains meaningful equity when called and can win the blinds without a showdown. Even against pocket kings, A6 suited has approximately 30% equity.

The opponent happened to hold one of the strongest hands in the calling range, and I was eliminated. That outcome did not change the quality of the decision. Poker continually offers the opportunity to make a correct choice and receive an unpleasant result, which is one of its more persistent educational features.

Nothing Left to Chance Except Chance

The deep run did not depend on the deck cooperating, which was fortunate because it showed no particular interest in doing so.

I had prepared from the estimation qualifier onwards, changed my training when the expected format changed, studied satellite and short-stack theory, watched the American Showdown, reviewed previous competitor accounts and built an initial model of the field.

I estimated the stack required to reach the final two tables, prepared the range framework, rehearsed a long live session, visited the venue, planned the physical preparation and carried backup food and caffeine.

Each piece served the same purpose: reducing the number of avoidable ways the opportunity could be lost.

Preparation cannot guarantee the favourable branch of a probabilistic event. It can make the adverse branch playable.

I reached the final two tables, finished inside the top 18 and earned the interview invitation. The prize money would have been welcome, but the interview was the reason I was there.

The experience confirmed something I already suspected about myself. I do not need to believe that preparation defeats variance, and I am perfectly capable of recognising when another competitor has done the work and played well. I do need to know that I gave the opportunity everything available to me.

Roy Jones Jr. once expressed that competitive intention rather neatly:

You will not win because I will not lose.