Aleatoric FC

Claude Shannon — the father of information theory — once asked Edward Thorp on their way to the casino: "What makes you tick?" Two of the greatest minds of the 20th century, heading to beat roulette with a homemade wearable computer. This actually happened.

Shannon Entropy: Measuring Uncertainty in Information H(X) = - ∑ P(xᵢ) log P(xᵢ) This is the legendary formula by Claude Elwood Shannon (1916–2001); the father of Information Theory. Entropy quantifies how much uncertainty (or average information) is contained in the outcome of a random variable X. The more unpredictable the outcomes, the higher the entropy. From data compression and cryptography to AI and communications; this concept powers the digital world.

Richard Feynman explaining atoms and stars in this ‘Fun to Imagine’ BBC series, recorded at Feynman's home in Altadena, California, ca. 1983 🧠 - 📹 BBC Archive

Essential reading to understand the variance of losing drawdowns. BettingIsCool provides the wonderful scripts. My book provides the background maths.

Very random thought: wouldn’t “AGI” commoditize “AGI” the moment any company releases it to the public? What is the point of getting first to a technology that will render the very creator useless and undifferentiated? Most of the argument on AI spending relies on the good guy getting there before the bad guy. And it seems very counterintuitive to me. I woul love your thoughts on this @sharps_research @12Xpert

💯

@aakashgupta @ylecun The GPUs and ScaleAI, which they thought to be a better bet that @ylecun From the People who brought us the Metaverse… it could not get better!

Yann LeCun closed $1.03B for AMI Labs on March 10. Three days later, this paper dropped from his NYU collaborators. 15M parameters. Single GPU. A few hours of training. LeWorldModel is the first JEPA that trains end-to-end from raw pixels. Two loss terms: predict the next embedding, keep the latent space Gaussian. Previous JEPAs needed exponential moving averages or pretrained encoders to avoid representation collapse. LeWM doesn't. Six hyperparameters down to one. The numbers are the story. Foundation-model-based world models require hundreds of millions of parameters and serious compute to plan a control task. LeWM plans up to 48x faster while staying competitive on 2D and 3D benchmarks. The whole thing fits on a laptop GPU. Look at the trajectory. Yann announced his Meta departure in November 2025 after 12 years and called founding FAIR his "proudest non-technical accomplishment." On March 10, 2026, AMI Labs closed the largest seed round in European history at a $3.5B pre-money valuation. Bezos, Nvidia, Samsung, and Toyota all wrote checks. Three days later: a paper showing that JEPA-from-pixels is no longer fragile and no longer compute-heavy. The engineering scaffolding that made it look like an academic curiosity is gone. The authors sit at Mila, NYU, Samsung SAIL, and Brown. None at Meta. Yann's bet was that the path to machine intelligence runs through world models, not language models. He left a public company to build it. Each JEPA paper from his network resets the assumed cost structure for that bet. This one makes world modeling laptop-cheap. Meta still has the GPUs. The architecture left.

La habilidad más infravalorada del inverso medio: La paciencia 📶

@WiLLyTfE1 @NeilMac555 Okay, so what you are saying is that those movements, not always, but sometimes create opportunities. Because the market overreacts. Your algorithm filters those ocasiona for profit. Is that it?

%CLV is a very accurate metric for comparing your long-term %YIELD. But when you’re investing just a few minutes before the markets close, it won’t matter much. My bot, for example, posts 3 minutes before the kickoff time. As I told you, for this type of approach, market movements, market patterns, and robust, consistent parameter filtering are the most important factors.

Barca 1.81 to win El Clasico this weekend

@12Xpert @DivinelyDesined Do you believe there is a healthy way of combining both believe systems (for a person, not society)? I know it might be a stupid question.

@AleatoricFC @DivinelyDesined Pretty much. It's all a comfort blanket for folk that suffer anxiety about uncertainty. Science's currency is uncertainty. Faith's currency is absolutism. The first offers an endless journey into curiosity, the second a dead end cul-de-sac to boredom.

The incredible DNA Repair System DNA - and Life - cannot exist without it. It requires 6-7 sub-systems working together + a massive complex network of 130–200+ unique proteins. Simpler versions don't work. Only Intelligence designs systems like this.

@ChapinPatrick In 2018, your odds of dying from COVID-19 were zero 🤔

@sharps_research @jomatech was so on point! Founder mode is back!

@WiLLyTfE1 @NeilMac555 Do I understand correctly that the market is inefficient at processing information? It overcorrects , creating opportunity. Is that correct? The reason I ask is that if this is true, then CLV loses value as a metric.

I started in September 2023 by tracking market movements in sufficiently liquid markets (<5% VIG) during the 80 minutes leading up to kickoff. I can’t speak to what I haven’t analyzed (i.e., matches with lower liquidity and movements that occur more than 80 minutes in advance). What I’ve learned over these nearly three years is that if you do steam betting, you’ll have to do it with soft bookies (to increase your edge). Steam betting at sharp bookies won’t yield long-term profits (if you simply bet on a market that has, for example, seen a 5% drop in odds). However, there are other ways to become profitable by following market movements during the 80 minutes before kickoff (more liquidity and information enter the markets). It’s not easy; I’ve made many mistakes over this time. You have to take into account opening odds, market patterns (each market behaves differently), and find robust, consistent filters over time (ones that perform well in backtesting but also generalize well in the future). Overfitting is inevitable; the key is that, within the complexity, the filtering should be as simple as possible and follow the most logical approach. Best regards.

@12Xpert @DivinelyDesined Is the creator humanity’s way of copping with the fact that we are not in control? Is it easier to accept that there is an omnipotent version of ourselves that loves us, that is the creator of the universe?

@DivinelyDesined It's funny how for a long time creationists denied science and even killed scientists for their work. How Christian of them. Nowadays they can't do that, so fraudulently use science to defend their position. I guess at least that's better than burning people they disagree with.

@PlusEVAnalytics @FezzikSports Is this where your post of Sharpe Ratios was heading too?

Sizing matters. If the typical sports bet is $100 @ -5% EV and the typical lottery bet is $5 @ -50% EV, the lottery is better "value". Of course sizing varies from person to person, but among those who play both sports and lotto it would be very unusual to see same sizing on both. Sizing is generally an inverse function of variance. We need a unified theory of EV, scaled for sizing, that can apply to straight bets, parlays, roulette, lotteries etc. Maybe I'll come up with one.

People are OUTRAGED that almost everyone loses betting sports at a -5% ish house edge on straight bets. Yet, the Governments keeps hyping the LOTTERY spending millions on ads, with a 50% house edge.

@NeilMac555 @WiLLyTfE1 @WiLLyTfE1 what is the verdict?

@AleatoricFC Talk to @WiLLyTfE1 he has an interesting take on these moves and has been profitable lately

@NeilMac555 Yes, and it is valuable info. Really like what you do. And I am trying to understand if that movement creates an opportunity or not. Looking at what happens historically in such occasions might help answer that.

@AleatoricFC It's nuanced, but what I was showing in the above post was the market reaction to the news

@NeilMac555 I get that, but do such movements end up predicting what actually happens?

@AleatoricFC The market moved in Barcelona's favour because of the drama that is unfolding at Real Madrid. Valverde and Tchouaméni will be added to the absentee list

@NeilMac555 If you have all historical movements (filtering by size, the biggest ones), and you have the game results, you could simply calculate an accuracy score treating the movements as predictors. Or calculate the correlation of outcome and movement.

@AleatoricFC I can’t see if the movement actually correlates with reality. Explain that

@NeilMac555 I have been following your posts on market movements for a bit now. From the few samples, I can’t see if the movement actually correlates with reality. Is that the case in your experience? If so, it would make the case for technical analysis type of betting I guess?

update

A WSJ analysis found 67% of Polymarket profits goes to just 0.1% of accounts, while most traders are in the red. Most Kalshi users also lose money. on.wsj.com/4td9UPy