About Paradis

I write here as Paradis.

The name is a nod to Bryce Paradis, one of the old Limit Hold’em crushers whose CardRunners videos changed the way I thought about poker. By the time I found those videos, poker had already taken hold of me, although they helped show me how much further the game could be taken once instinct was joined by serious technical study.

Poker

The first book that properly hooked me was Small Stakes Hold’em by David Sklansky and Ed Miller. It introduced me to the mathematical structure underneath poker and moved the game well beyond guessing somebody’s hand, studying their face or listening to an old live pro explain that he had “put him on ace-king” because of a feeling.

There were reasons behind the decisions: pot odds, equity, protection, value, free cards and thin advantages that mattered because they repeated. Once I could see structure underneath the noise, I wanted to understand it properly.

I started in live Limit Hold’em, a game of fixed bets, repeated decisions and relatively little theatrical nonsense. There were no dramatic river overbets or solver-approved betting trees. There were streets, discipline, boredom, ego and small advantages that had to be pressed repeatedly until they became money.

I played live Limit Hold’em up to 10/20, and for a while poker paid my rent. That sounds cleaner than the underlying experience.

Live poker was a strange world. Sometimes Russian oligarchs splashed money around as though it had lost all meaning. Sometimes people were wheeled out of bathrooms after overdosing. On one occasion, a PLO player completely lost his mind, grabbed his opponent’s chips, threw them across the room and screamed, “You want free money? Any of you want free money? Free money for everyone!” while security came running from every direction.

God bless PLO players.

Most of the work was considerably less dramatic. The hardest part was often folding.

Before earphones were allowed, you could spend hours folding every hand before the flop while the table laughed, drank, needled one another and supplied every emotional reason to become involved. You would eventually leave feeling satisfied because you knew you had made the correct decisions, even though almost nothing visible had happened.

That was one of the first useful lessons poker taught me. Discipline does not always feel intense or heroic. Much of the time it is excruciatingly boring, and the best decision can look indistinguishable from doing nothing.

Back when the room still permitted it, I occasionally played 24-hour sessions. They were ugly, exhausting and fuelled by more casino coffee than any functioning digestive system should be expected to tolerate. When the game remained good and I could still think clearly, the decision was straightforward: remain disciplined and continue making better decisions than the people giving their money away.

Edges are not always clever. Sometimes the edge comes from being willing to make the correct boring decision longer than everyone else.

When the live Limit Hold’em games dried up, I moved online. The environment was faster, sharper, less forgiving and much more technical. Improvement was mandatory because the games had no interest in preserving anybody’s self-image.

I eventually played online Limit Hold’em up to 15/30, became a PokerStars Supernova and earned Full Tilt Black Card status. Volume mattered. Rake mattered. Table selection mattered. Tilt control mattered. The instincts developed in live poker still helped, although they could no longer carry the full weight of the decision-making process.

This was where Bryce Paradis and the training-site era became especially important to me. Poker could be understood through frequencies, equilibrium, exploitation, indifference and game theory. The game was deeper than I had realised, which naturally made me want to spend even more time on it.

There was always another layer. First you learned to stop playing trash. Then came pot odds, equity and ranges, followed by the realisation that your opponent understood ranges and knew that you understood them as well. Before long, the game had become a hall of mirrors with money in the middle.

Underneath all of it remained a mathematical problem waiting to be understood. That is still the kind of problem I like.

The Long Route into Data Science

I also spent time as a medic in the Army Reserve and came first in the Eastern Region Camp Med Assist course for 5 Brigade. I mention it because it was another environment where competence came from repetition, standards and doing the work properly regardless of how I happened to feel on the day.

Later, I found my way into data science. In hindsight, the direction makes sense. I already liked messy problems where the answer was not supplied cleanly, and data science gave me stronger tools for investigating them: Python, statistics, machine learning, validation, simulation, imperfect datasets and the discipline of turning a suspicion into something testable.

I eventually became a senior data scientist at a leading data consulting firm without following the conventional route into the profession. No degree opened the door for me, so I studied after work, filled the gaps I could identify, rebuilt foundations where necessary and kept moving towards the problems I wanted to solve.

The process was lengthy and frequently unglamorous. Most serious learning seems to involve more evenings alone with a textbook than the brochures imply.

The MITx MicroMasters in Statistics and Data Science came from the same instinct that had driven my poker study. The certificate had value, although the foundations were the real reason I enrolled. I wanted to understand probability, statistics and machine learning properly, rather than accumulate disconnected techniques and trust that they would somehow hold together when the problem became difficult.

My starting point and the absence of a clean route were largely outside my control. The direction, the standard of the work and the number of hours I was prepared to put into it were available to me, so that was where I concentrated my effort.

Grinding has limits, and effort cannot rescue a bad model or replace ability. It remains one variable I can control more reliably than most.

Why MathematicalEV Exists

MathematicalEV grew from the same instinct. It is where I put the questions I cannot leave alone.

A shuffling machine that everyone treats as random. A poker model that fails at the edge. A chess position where studying pawn tension reveals something about deep-stacked poker. A model that works beautifully until somebody asks the wrong question. A game where the interesting problem includes what the structure forces somebody to do, rather than only what they currently hold.

I like mechanisms. I want to know what is happening underneath, particularly when a claim sounds suspiciously clean.

“It is random.” Fine, what process produced the output?

“The model works.” Under which conditions does it stop working?

“That is just how the system works.” Why?

That is why the site moves between poker, shufflers, probability, decision-making, data science, chess and whatever other strange corner catches my attention. These subjects have different rules and specialist knowledge, but they frequently involve the same deeper problems: decisions made with incomplete information inside systems that are messier than people prefer to admit.

The pawn-tension article is a good example. While studying chess, I encountered an explanation of a beginner releasing tension in the centre. That idea exposed a leak in my deep-stacked poker thinking that years of poker vocabulary had failed to make obvious. The value extended beyond one chess position or one poker hand. Studying structure in one domain had made hidden structure visible in another.

The ShuffleMaster project began with an ordinary-looking machine described using one reassuring word: random. That word did not answer the question for me. I wanted to know what physical process generated the cards, which assumptions were hidden inside it and whether its output behaved like the probability models people casually treated as equivalent.

The question became a working simulator, and the simulator became a much larger investigation than I initially expected. The site records the project as it develops, including the assumptions, failed approaches, corrections, code and the points where an effect begins to look real.

Over time, I realised that the subjects were only part of the story. Each article also reveals something about how I approach a problem: what attracts my attention, where I suspect an edge, what makes me distrust an easy answer, how I respond when a theory fails and how far I am prepared to follow a question once it takes hold.

I do not want the writing polished until every trace of the original thought has disappeared. Some pieces will be technical, others speculative, and some will contain mistakes that become useful later. Following an idea honestly requires preserving enough of the process to show how the conclusion was reached and why it changed.

The point is to keep pulling on the thread until it collapses or becomes something worth keeping.

Away from the Table

Away from the table and the screen, I live with my partner and our black German Shepherd, Panda. I cook, lift, run, study, play poker and spend too much time wondering whether an ordinary-looking system is concealing something interesting.

The phrase I keep returning to is:

Mollitia exitium est: softness is ruin.

I do not take that as an instruction to make life permanently grim. I enjoy good food, good wine, dogs, games, jokes and good conversations far too much for that interpretation to survive contact with reality.

For me, it means that comfort cannot be allowed to make the decision.

That is the standard I am trying to hold myself to here.