9/29/2026

Risk Premium Research · Themis

Fifteen CEOs, scored on what was theirs

An attribution-adjusted assessment of fifteen anonymized chief executives across industries. Each rating strips out inherited strategy, cyclical tailwinds, one-time items and base-effect recovery before it credits the leader.

Scores as of 29 May 2026Market data refreshed to 28 Sep 2026Published 29 Sep 2026

Composite comparison

The ranking

Ranked by raw composite, the simple average of thirteen dimension scores. The raw composite answers "how well led is this company". Delta and TSR residual answer the harder question: how much did the leader add beyond the hand they were dealt. Click a column heading to re-sort.

	Case	Tenure					TSR residual

8 to 106 to 7.94 to 5.9below 4Delta = Raw minus Baseline. Founder baselines are n/a by convention.

Baseline: the company at handoverRaw composite: the leader's scorecardThe line between them is the delta.

How to read it

Two measures of CEO contribution

The most common analytical error is to confuse the baseline (the hand dealt) with expected performance (what a peer-median CEO would deliver with that hand, in that cycle). They are different concepts and produce different measurements.

Delta from baseline

Raw composite − Baseline composite

Asks whether the company's qualitative state improved during the tenure. A direction-of-travel check. It says nothing about what a competent replacement would have done.

TSR residual vs expected

Actual TSR − (Peer-median TSR + difficulty adjustment)

Asks whether the CEO beat a peer-median replacement holding the same hand over the same window. This is the value-creation measurement. Total return includes reinvested dividends; buybacks are judged under capital allocation, not added to TSR.

What changed since May

The September refresh

For the eleven tenures still running, we pulled verified total returns from 29 May to 28 Sep 2026 for the company, its peer set and its index, then checked whether any TSR band, CEO status or consensus view had changed. Four tenures had already closed. No point score moved. Three scores now carry a range, and one case was rebuilt from scratch.

#	Case	Status	What the refresh found	Consensus (analysts)

Case files

Fifteen scorecards

Each dossier shows the setup, the key metrics, the verdict and the thirteen-dimension scorecard. Scores of N/A are excluded from the averages. Value weighting doubles Strategic Vision, Capital Allocation, Operational Execution and TSR, the four dimensions that drive durable value.

Closing observations

What the fifteen cases show

Originators sit on top, hired or founder

The top four all originated the thesis that now defines the company, and one of them is a hired CEO. Stewards of fortresses come next, then executors of inherited plans, then leaders ousted or churned under cause. The spread runs 2.7 to 8.8 without collapsing to a middle.

A high baseline is a test, not a credit

Payments and aerospace both inherited near-duopoly franchises. One turned that hand into a strongly positive residual; the other into a deeply negative one. Seven of the eight negative deltas belong to leaders who inherited baselines of 6.5 or higher.

Where the framework parts with the Street

On the main aggregator, consensus rates every live case Buy or better, including aerospace at Strong Buy. The framework is harsher on competent operators of inherited premium franchises. Sell-side prices the stock; the framework grades the leader. Both can be right at once.

Method and provenance

How the numbers were built

The framework

Four adjustment disciplines: strip one-time items, separate cyclical from structural, separate inherited from originated, separate input metrics from output targets.
Four symmetry gates: equal-depth evidence sheets, a challenger pass on every score of 7 or above, a peer-median counterfactual, and an insider pushback test.
Baseline scorecard: five factors scored at handover (franchise, balance sheet, culture, strategic clarity, cycle position).
Every score is anchored to a hard metric and threshold, or to observable components combined by a stated rule.
Hard rules: an ousted CEO caps at 3 on Board & Governance; a 9 or 10 on Strategic Vision needs two dated, pre-consensus decisions.

This refresh

Prices: exchange closes on 29 May and 28 Sep 2026, cross-checked on a second source where available. Total return = (end price + dividends with ex-dates in the window) ÷ start price − 1, in local currency.
Consensus ratings and average targets: aggregator pages dated 11 to 29 Sep 2026.
All composites were recomputed in code from the dimension scores. Four differed from the values recorded in May by 0.2 to 0.4 points, and this page uses the recomputed values. Enterprise software rises from 7.8 to 8.1. Commercial aerospace falls from 3.5 to 3.1. Payments rises from 7.0 to 7.2. Premium footwear rises from 5.4 to 5.6. The rank order is unchanged apart from the payments and gold-mining swap.
Two pairs are ties on raw composite (5.92 and 5.58) and should not be read as ranked.

The rebuilt case

The mass-retail case was rebuilt from the verified public record. Its earlier scorecard drew on evidence from an unrelated case, so it was replaced rather than refreshed. Every figure in the new version is sourced to company filings, releases and exchange prices: tenure total return against the two main peers and the broad market, the operating margin at the 2021 peak and at exit, and the timeline of the two margin goals. Its baseline of 4.8 is a new judgment-anchored score.

Known limits

The framework scores observable performance. It cannot detect undisclosed fraud or concealed misconduct.
Four refresh inputs are unverified or weak: one financial-sector index start price (undated source), one medical-device peer start price (not found), one peer end price from a low-volume snapshot, and a sell-side split for two names.
Residuals over short tenures are dominated by noise. Treat them as point estimates with wide error bars.

9/24/2026

In memoriam: Armindo Manuel Santos Pinto (1941-2026)

- Pedro

Read on Substack

7/12/2026

My Model Learned a New Trick. It Immediately Changed Its Mind About Who Wins.

The quarter-finals are done, the field is down to four, and this week I did something I had been putting off: I let the model learn from the knockouts. Until now it froze every team’s strength at the end of the group stage and just tracked the bracket. Sensible, but half-blind. It could see who survived. It could not see how they were playing.

So I fed the knockout results back into the ratings. Now it knows the difference between winning well and just winning. And the instant it could tell them apart, it moved the favourite off Argentina.

Here is where it lands, what the upgrade changed, and the one call where I now disagree with my own code more sharply than at any point in this tournament.

Subscribe now

The state of the race

Conditioned on every result through the quarter-finals, and now with knockout form baked into the ratings, the title picture:

 Spain 30.6% ·  Argentina 27.9% ·  France 24.8% · gbeng England 16.7%

Forecast Wc2026 V10

612KB ∙ PDF file

Download

Four teams left, so those four hold the entire championship probability between them. The single most-likely final the model produces is still Spain vs Argentina, and its predicted champion is now Spain. Read the top two lines carefully, because the order just changed, and it changed for a reason.

The setup, for anyone new here

The engine is an ensemble of two models: a Dixon-Coles goals model and a World Football Elo rating, blended at the scoreline level and run through 10,001 Monte Carlo simulations from here to the final. It does not get nervous. It does not have a favourite nephew. It does maths, and this week it did some new maths.

The upgrade: I let the model watch the football

The old version had a blind spot I had been honest about since the group stage. It locked each team’s strength when the groups ended and then simply cascaded the bracket forward. That meant a team could win three knockout ties by a single goal, riding its luck, and the model would treat it exactly as it treated a team dismantling opponents 3-0. Survival and dominance looked identical on the page.

The fix was to feed the knockout results back into the ratings, so the model now updates on how teams are actually playing, not just on who is still alive. It is a small change in code and a large change in worldview. The moment the model could separate winning well from merely winning, it re-ranked the favourites. That is the whole story of this update.

What the model now sees

 Argentina demoted, from 33.6% down to 27.9%. Three knockout wins, yes, but it conceded in every one of them. It survived Egypt and it survived Switzerland rather than dismissing either. The old model rewarded the clean bracket path. The new model looks at the performances underneath and marks them down. The bracket had been flattering it.

 Spain surged to number one, from 25.7% up to 30.6%. It took Portugal apart and then took Belgium apart, beating strong teams convincingly. That is what actually moves a rating once the model is allowed to watch. Spain did not back into this. It earned it on the pitch.

A quiet flex I have earned the right to: my pre-tournament top two, Spain and Argentina, are both still standing, and the entire final four is exactly the calibre the frozen forecast flagged before a ball was kicked.

The semis

 France v Spain  · gbeng England v Argentina 

A France v England final is very much alive, and if you have read these updates since the group stage you know exactly which of those two names I have been shouting about for weeks.

Where I overrule the machine, and it is not close

Here is the honest part, and it has sharpened since the last edition, so let me say it plainly.

Two rounds ago I said France and England, both, were operating on a different frequency. The bracket has since answered part of that for me: England drew Argentina, France drew Spain, and my two-team thesis has to survive contact with two very hard semi-finals. So I will narrow it and commit. The model has France third. I have France first, by a distance, and I will die on this hill.

Under the hood

For the curious: the two models are blended at the scoreline level and the remaining tournament is simulated 10,001 times, now conditioned not only on every actual result but on updated strength ratings that absorb knockout form. The rigour from earlier editions still holds. FIFA’s 2026 tiebreaker puts head-to-head above overall goal difference, a break from every previous World Cup, and it is coded properly rather than approximated. The knockout third-place allocation follows FIFA’s official deterministic table. A forecast is only as trustworthy as the plumbing beneath it, and this week the plumbing got an upgrade.

What is next

The semi-finals are loading. France v Spain, England v Argentina, no redraws, no second chances. The model gets re-run after every round, and I will keep publishing where it lands, and where, like now, I loudly disagree with it.

Model says Spain. My eyes say France. We are about to find out who was right.

The Champion Predictor is a quantitative model built for research and curiosity, not betting advice. For round-by-round updates through to the final, subscribe. The next one lands after the semi-finals.

from Risk Premium https://ift.tt/MkmpbSu
via IFTTT

7/08/2026

My Model Now Says Argentina by a Mile. My Eyes Say France and England.

The Round of 16 is done, the field is down to eight, and the model has re-run the whole tournament on the wreckage. Two title contenders gone from the same half of the bracket. The frozen pre-tournament forecast still standing on five of the eight quarter-finalists. And a champion probability that just lurched in one direction hard enough to make me suspicious of it.

Here is where it lands, who the penalty spot just swallowed, and the one call where I still refuse to agree with my own code

The state of the race

Conditioned on every result through the Round of 16, the title picture:

 Argentina 29.2% ·  Spain 21.4% ·  France 19.2% · gbeng England 13.6% ·  Morocco 6.5%

The top four now hold 83% of the championship probability between them. The single most-likely final the model produces is still Spain vs Argentina. Read the top line carefully, because it moved for a reason that has nothing to do with Argentina playing well

Forecast Wc2026

571KB ∙ PDF file

Download

The setup, for anyone new here

The part I am still proudest of

Anyone can look clever editing predictions after the goals go in. The honest test is the version you lock the night before the opener and never touch again.

That frozen forecast called five of the eight teams now in the quarter-finals, sight unseen: Argentina, Spain, France, England, Belgium. The three it missed are exactly the tournament’s fairytales: Morocco, Norway and Switzerland. A model built on 49,390 historical matches was never going to price a fairytale. That is not a bug. That is the fairytale doing its job.

For the record, the frozen forecast also nailed 12 group winners out of 12 before a ball was kicked. I will be mentioning this at dinner parties until roughly 2030.

Subscribe now

The fallen

The Round of 16 sent two genuine contenders to the airport, and it took them from the same corner of the bracket.

 Brazil, whom the model gave a 10.3% title shot at the last update, walked into Norway and lost 1-2.  Colombia, a 5.8% side going in, went out to Switzerland on penalties.

The penalty spot remains the place forecasts go to be humbled. A shootout is a coin flip in expensive boots, and no model that respects the data pretends otherwise. This time it did not just take two contenders. It gutted an entire half of the draw, which is the single most important fact in this update.

What moved since Version 7

Compared with the forecast I last published at the end of the Round of 32, the knockouts did what good information should do. They sharpened the picture, and in one case rewrote a team’s odds without that team doing anything at all.

 Argentina leapt from 21.8% to 29.2%, a +7.4 jump. It did not surge because Argentina played well. It needed a nervy 3-2 over Egypt just to get here. It surged because the two teams that could have stopped it, Brazil and Colombia, both went home on the other side of the bracket. The model is not pricing Argentina’s football. It is pricing Argentina’s suddenly empty path.

gbeng England is the quiet mover of the round, up +6.6 from 7.0% to 13.6% and climbing.  Spain also jumped hard, +5.8 from 15.6% to 21.4%, and now sits clear in second.  France barely budged, +0.8 from 18.4% to 19.2%, because unlike Argentina it is stuck in the hard half with Spain, Morocco and Belgium. Your odds are hostage to the draw as much as to your own form. Easy bracket and best team are graded on very different rubrics, and the bracket is the rubric that pays out.

Where I overrule the machine

Here is the honest part, and it has shifted since the last edition, so returning readers deserve me saying so plainly.

Last time I overruled the model in favour of France, alone. I still believe in France. But watching this round, I now think there are two teams operating on a different frequency, not one. France and England, both. Sharper, deeper, meaner from front to back, peaking at exactly the right moment. To my eye that is your final. France vs England. Book it.

And Argentina? The model’s 29% darling looks, to me, the shakiest favourite of the lot. A 3-2 wobble past Egypt is not the swagger of a 29% team. It is a side being carried by a collapsing draw. The model rewards the path. I am watching the performances, and the performances are flashing yellow.

One of us is going to look very silly when this is over, and I am suspiciously comfortable betting against the algorithm’s pet. A forecast you only ever nod along with is not worth much. The value is in the disagreement, and in being honest about which side of it you are standing on.

Under the hood

For the curious: the two models are blended at the scoreline level and the remaining tournament is simulated 10,001 times, conditioned on every actual result so far. The same two acts of rigour from the last edition still hold. FIFA’s 2026 tiebreaker puts head-to-head above overall goal difference, a break from every previous World Cup, and it is coded properly rather than approximated. The knockout third-place allocation follows FIFA’s official deterministic table. A forecast is only as trustworthy as the plumbing beneath it.

What is next

The quarter-finals are loading. The bracket is fixed, no redraws. The model gets re-run after every round, and I will keep publishing where it lands, and where, like now, I quietly disagree with it.

Model says Argentina. My eyes say France and England. We are about to find out who was right.

The Champion Predictor is a quantitative model built for research and curiosity, not betting advice. For round-by-round updates through the knockouts, subscribe. The next one lands after the quarter-finals.

from Risk Premium https://ift.tt/I5BFKti
via IFTTT

7/05/2026

Eight Novels for the Summer

The Economist ran its summer fiction list on June 26. Eight titles. A murder mystery, two Irish and Italian family sagas, a Bangladeshi rebellion, a Mississippi Depression story, a slim Ben Lerner puzzle box, and a Patchett reunion. All eight are new this year. Below is what each one is and why it earns a place on the summer stack, with nothing given away.

The Calamity Club, by Kathryn Stockett

Seventeen years after The Help, Stockett’s second novel. Oxford, Mississippi, 1933. Prohibition is fading, the Depression is biting. An eleven-year-old orphan, an unmarried woman down from the city, and a stranger with little left to lose cross paths. What follows is a story about women taking back agency in a place built to deny it. Two narrators, two timelines, one long-awaited return.

The Kindness of Strangers, by Emma Garman

A debut, and a good one. London, 1953. A boarding house in Chelsea, a widow who runs it, and a stranger who arrives one foggy night and unsettles everyone under the roof. The book opens with a man dying on the drawing room floor. Everything after works backward and forward from that. Compared to Christie and Kate Atkinson, but the voice is its own. Postwar London rendered in smoke and food shortages. The Economist called it sharp and full of period detail.

Land, by Maggie O’Farrell

Her tenth novel, and her most ambitious. Ireland, 1865, a decade after the Great Hunger. A mapmaker working for the British Ordnance Survey has an epiphany at an ancient spring and abandons the official map for an unofficial one, told in the native language. A family saga that stretches from one life span to the age of the land itself. Colonization, displacement, folklore. The kind of book you live inside rather than read.

Other People’s Children, by Ben Faccini

Faccini’s third novel, and his first in nearly twenty years. A man named Tommaso holds his life together: a job that sends him abroad, a partner with two unruly sons, and an aging Italian grandmother, Alma, whose memories are turning strange. The past reaches back to the Italian resistance in the Second World War and forward into the chaos of modern London. A quiet, precise book about inheritance and the long shadow of what families keep hidden.

The Things We Never Say, by Elizabeth Strout

A new standalone from the author of Olive Kitteridge and the Lucy Barton books. Strout does what Strout does: ordinary people, subtle emotional weather, the things left unsaid between them. If you have read her before, you know the register. If you have not, this is a fair place to start.

Transcription, by Ben Lerner

Slim. Formally unstable, as the New Yorker put it. Lerner’s follow-up to 10:04 and The Topeka School, a novel that blends fiction, memoir, and essay, shot through with humor and anxiety. A meditation on memory, fatherhood, and the devices we use to store or erase experience in a digital age. It reads fast and settles slow. Winner of the Orwell Prize.

Uprising, by Tahmima Anam

Inspired by the real women of Banishanta, Bangladesh. On a sinking island off the coast, a community of women live under a madam who was once sold into slavery herself. Their children narrate. When an educated young woman is forcibly brought to the island and refuses to yield, complacency turns to defiance. A fierce, slim, feminist novel of resistance and female power. Finalist for the Orwell Prize for Political Fiction.

Whistler, by Ann Patchett

Her tenth novel, and by several accounts her best. It opens in the Metropolitan Museum of Art, where a woman in her fifties notices an old man following her and her husband through the galleries. He is her former stepfather, whom she has not seen since she was nine. What binds them is a single winter day decades earlier and a story he once told her. Quieter than Bel Canto. A mystery of the heart rather than a body on the floor. Reconciliation, memory, and the small moments that define a life.

Eight books, one summer. Pick one, pick all of them. If you have read any, tell me where the rest of us should start.

from Risk Premium https://ift.tt/yf6Rno2
via IFTTT