https://www.playlikeaquant.com/all-experiments daily 0.75 2026-07-12 https://www.playlikeaquant.com/all-experiments/my-laptop-played-poker monthly 0.5 2026-07-12 https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/a728e1d7-b977-4112-ac88-f8d8049f2b78/Regret+Over+Training Experiment Archive - My Laptop Played 3.5 Million Hands of Poker Against Itself. Here’s the Bot it Became. - Make it stand out The whole project in one curve. “Regret” measures how much the bot wishes it had played differently. It spikes early as the bot discovers new ways to bet, then falls for three million hands straight as the strategy settles. By the end it has dropped to 0.076. https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/e0d324a8-d5cf-41c9-8233-5ade0527da43/Training+throughput+%28v2+vs.+v3+serial+vs.+parallel%29 Experiment Archive - My Laptop Played 3.5 Million Hands of Poker Against Itself. Here’s the Bot it Became. - Make it stand out Hands trained per second. The richer betting in v3 made each hand about three times more expensive to learn from, which is why parallelism stopped being a nice-to-have. Spreading the work across six cores took the bot from eight hands a second back up to nearly eighty. https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/102f4410-f35a-4641-8e58-e4016e02f6c5/Raise+sizes+and+information+sets Experiment Archive - My Laptop Played 3.5 Million Hands of Poker Against Itself. Here’s the Bot it Became. - Make it stand out Two ways v3’s world got bigger: one more bet size, and twice as many possible situations to learn. The number of situations the bot can imagine doubled; the number it ever actually meets in real play turned out to be only about a thousand. https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/04dfecda-a191-4d65-a49b-2dac9097db93/One-third-pot+experiment Experiment Archive - My Laptop Played 3.5 Million Hands of Poker Against Itself. Here’s the Bot it Became. - Make it stand out The fifth bet size that did not make it. Adding a one-third-pot option made betting sequences so long that the bot trained roughly five times slower for no real gain in strength. https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/1e671148-05d2-4215-97f9-2e4f0d3396c9/Distribution+of+raise+sizes Experiment Archive - My Laptop Played 3.5 Million Hands of Poker Against Itself. Here’s the Bot it Became. - Make it stand out Every raise the trained bot makes, sorted by size. No single size dominates. This is the one chart that convinced me the “feel like a person” pass had worked: it is neither a calling station nor a maniac who only shoves. https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/76e4f011-a43b-4672-8bb4-9be8451a45c8/Preflop+strategy+by+hand+strength Experiment Archive - My Laptop Played 3.5 Million Hands of Poker Against Itself. Here’s the Bot it Became. - Make it stand out Before the flop, by hand strength and seat. Weak hands on the left are folded; strong hands on the right are played aggressively but with varied sizing. The bot is noticeably more willing to put chips in from the big blind, where it has already posted money in the pot. https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/211d80d8-2527-43d5-8568-a7a1c0eaca5a/Performance+against+different+opponents Experiment Archive - My Laptop Played 3.5 Million Hands of Poker Against Itself. Here’s the Bot it Became. - Make it stand out Chips won per hand against three opponents over the full run. The bot beats the random and push-or-fold players comfortably. Against a pure calling station (red) it slides into the red and stays there, and that is not a bug. https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/4c9e52f8-5ce1-4ea1-a38f-b3aaec70231d/Poker7.1+-+Stats.png Experiment Archive - My Laptop Played 3.5 Million Hands of Poker Against Itself. Here’s the Bot it Became. - Make it stand out Whatever it is, the way you tell your story online can make all the difference. https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/83cfa460-5087-471c-8acf-56a2b21f80a0/Training+throughput+over+time Experiment Archive - My Laptop Played 3.5 Million Hands of Poker Against Itself. Here’s the Bot it Became. - Make it stand out Hands per second across the full run. No slow leak, no thermal cliff, just a flat line for twelve hours. Boring, in the way you want a long compute job to be boring. https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/f822281e-be4a-4ed4-ab74-49a292550bd8/Information+sets+reached Experiment Archive - My Laptop Played 3.5 Million Hands of Poker Against Itself. Here’s the Bot it Became. - Make it stand out How many distinct situations the bot had encountered, over the run. It saturates almost immediately near a thousand and never approaches the 3,200 it could theoretically reach. The bot learns the poker that actually happens, not the poker that could. https://www.playlikeaquant.com/all-experiments/neural-network-for-golf-shots monthly 0.5 2026-07-12 https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/22ca895d-4ae3-4836-badf-bfbd7901b174/Three+charts+overview Experiment Archive - I Built a Neural Network for Golf Shots. It Found the Sweet Spot. - Make it stand out Left: Nearly 2,000 real driver shots measured by a TrackMan launch monitor. Faster swings generally go farther, but shots at the same speed still vary by more than 60 yards. Middle: The neural network predicts unseen shots more accurately than either linear baseline. Right: The network independently learned the cost of missing the center of the clubface—a relationship a straight-line model can't capture. https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/b1c472ea-c5cb-414d-a9db-405f269ade1d/Straight-line+model+diagram Experiment Archive - I Built a Neural Network for Golf Shots. It Found the Sweet Spot. - Make it stand out The fitted straight-line model, drawn as what it is: a single neuron evaluating the dataset's typical shot. Each input is multiplied by a learned weight (orange), summed with a learned bias, and the total is the predicted carry. The weights are already little discoveries: about +3.6 yards per mph of club speed, and almost nothing per millimeter of strike location, because no single straight-line slope can describe what strike location actually does. https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/59a25ecc-d183-4e39-a902-8333d0e2b8c3/Toy+classification+example Experiment Archive - I Built a Neural Network for Golf Shots. It Found the Sweet Spot. - Make it stand out A two-input toy problem, shown because you can watch the mechanism with your own eyes. Left: the four dashed lines are the four hidden neurons; each one only knows how to split the plane in two. Right: the second layer re-weighs those four straight opinions into a single bent boundary that catches the ring. Nothing about circles was programmed in. My golf network does the same thing in six dimensions, where I can't draw it for you: the sweet spot is this ring, one dimension at a time. https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/27c671c3-403c-4b91-9287-37e89f6d7a5b/Learning+rate+illustration Experiment Archive - I Built a Neural Network for Golf Shots. It Found the Sweet Spot. - Make it stand out Three balls on the same loss landscape (darker blue is lower). Too small a step (η = 0.02) and the ball crawls into the nearest dip and stays there, a local minimum. About right (η = 0.15) and it strides into the deep valley. Too big (η = 1.5) and every step overshoots the slope it measured; the ball ricochets off the landscape entirely, which practitioners experience as their loss becoming an error message. https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/1234282e-222b-411b-ba98-75426c47821a/Training+loss+curve Experiment Archive - I Built a Neural Network for Golf Shots. It Found the Sweet Spot. - Make it stand out My golf network's real training run: wrongness after each pass over the shots, falling three-hundred-fold. The fast early drop is the ball finding the valley; the long tail is it settling toward the floor. https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/47dc2f9a-4615-4eb0-a817-415c16973a9d/Prediction+comparison Experiment Archive - I Built a Neural Network for Golf Shots. It Found the Sweet Spot. - Make it stand out The straight line and the network predicting unseen shots, guess against truth. A perfect model would sit on the dashed line. The network's cloud hugs it visibly tighter at every distance. https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/4dff90c7-6229-4ffe-b7ff-a61f711f0b9b/Clubface+surface Experiment Archive - I Built a Neural Network for Golf Shots. It Found the Sweet Spot. - Make it stand out The trained network's predicted carry for every strike point on the face, at the stock 100 mph delivery. A dome peaking at 243 yards within a few millimeters of face center, giving up roughly 27 yards by the far heel, about 18 by the toe edge, and the most of all (38) up in the high-toe corner. Trust the shape more than the altitude: against the closest matching real shots, the model's absolute yardages run optimistic in the middle of the speed range. The holes are honesty: face regions with almost no real shots, where the model has no business drawing confident numbers. Averaged over eight training runs. https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/d886ccfb-aece-4268-9339-f5fff924da8f/Strike-location+experiment Experiment Archive - I Built a Neural Network for Golf Shots. It Found the Sweet Spot. - Make it stand out Every dot is a real driver shot: how much farther or shorter it flew than its delivery predicted, plotted against strike location. Gold dots are binned averages, the dashed line is the best straight-line fit, and the orange curve is the strike-only neural network. https://www.playlikeaquant.com/all-experiments/admissions-regression-model monthly 0.5 2026-07-12 https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/18d7cd9c-c143-4692-982d-63dfb30c3e90/Eight-school+scatterplot+grid Experiment Archive - Testing a Simple Regression Model on College Admissions Data. - Make it stand out Each panel shows one school’s applicants from this dataset. Notice how cleanly the blue dots separate from the red X’s in some schools (JMU, Georgetown), and how scrambled they are in others (Penn State, Northeastern). That visual gap is the whole story. https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/0025dda2-a0e0-4af0-a69c-d6ab26c59b81/Stacked+bar+chart Experiment Archive - Testing a Simple Regression Model on College Admissions Data. - Make it stand out Each bar shows how much of an admissions decision at that school is statistically explained by GPA, test score, and decision type (early vs. regular). The gray block is everything the numbers can’t explain: essays, extracurriculars, recommendations, and the intangibles of holistic review. https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/a1aa3b8a-1055-4b22-a4c9-e16d60780966/JMU+v+Northwestern+Decision+Boundaries Experiment Archive - Testing a Simple Regression Model on College Admissions Data. - Make it stand out Left: JMU, where the dashed line cleanly separates blue (accepted) from red (denied). Right: Northeastern, where blue and red are scattered all over each other and no line can really separate them. Background shading shows the model’s predicted probability of admission: warm cream means low, cool blue means high. https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/faf6f686-2f84-4ef1-8bbd-835a4a3d3d52/Northeastern+High+SAT+Scatter+Plot Experiment Archive - Testing a Simple Regression Model on College Admissions Data. - Make it stand out Several Northeastern applicants with SATs of 1500+ were denied or waitlisted. The pattern is consistent with what some college counselors call yield protection, but I want to be clear that this is one possible explanation among several, and the sample here is small. https://www.playlikeaquant.com/all-experiments/category/Life+by+the+Numbers monthly 0.5 https://www.playlikeaquant.com/all-experiments/category/Golf monthly 0.5 https://www.playlikeaquant.com/all-experiments/category/Fun+with+Desmos monthly 0.5 https://www.playlikeaquant.com/all-experiments/category/Exploring+Pathways monthly 0.5 https://www.playlikeaquant.com/all-experiments/category/Admissions+Analysis monthly 0.5 https://www.playlikeaquant.com/all-experiments/category/Teen+Financial+Literacy monthly 0.5 https://www.playlikeaquant.com/all-experiments/category/Machine+Learning monthly 0.5 https://www.playlikeaquant.com/all-experiments/category/Stats+in+Action monthly 0.5 https://www.playlikeaquant.com/all-experiments/category/Game+Theory monthly 0.5 https://www.playlikeaquant.com/all-experiments/category/Quant+Experiments monthly 0.5 https://www.playlikeaquant.com/all-experiments/category/Interactive monthly 0.5 https://www.playlikeaquant.com/all-experiments/category/10+Minutes+with+a+Quant monthly 0.5 https://www.playlikeaquant.com/all-experiments/category/Sports+by+the+Numbers monthly 0.5 https://www.playlikeaquant.com/all-experiments/category/Quant+Questions monthly 0.5 https://www.playlikeaquant.com/all-experiments/category/The+Long+Game monthly 0.5 https://www.playlikeaquant.com/all-experiments/category/Market+Thinking monthly 0.5 https://www.playlikeaquant.com/home daily 1.0 2026-07-12 https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/028f3fac-381f-4fe5-b8e2-818ab3374409/Meet+Henry+Anderson https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/a3ff4881-25de-459a-80b6-112a3d134938/Henry+Anderson+Playing+Golf https://images.squarespace-cdn.com/content/v1/69bca85e190591600bbe558d/b607047d-05f5-45d9-a025-e9d189401e6d/Henry+Anderson+Playing+Lacrosse https://www.playlikeaquant.com/contact daily 0.75 2026-07-12