Kevin Coombes
krc77.bsky.social
Kevin Coombes
@krc77.bsky.social
Husband, father, grandfather, cyclist, tifoso, mathematician, biostatistician, bioinformatician. Reader of books, watcher of movies, listener to music, wanderer through museums. Appreciator of foreign travel.
In the USA, cycling fans have hardly ever been able to watch free, live cycling. Decades ago, all we saw was four Sundays of highlights of the TdF on ABC. In the Armstrong era, we got free daily coverage of TdF. Now I can cobble together 3 paid subscriptions to cover the season. Until HBO/MAX dies.
December 27, 2025 at 1:20 PM
Yes; Gianni Savio would be out there hustling full time to sell the rights to fill in those empty spaces.
December 23, 2025 at 12:16 PM
Actually, they look exactly like the chocolate sambuca cookie/biscuit recipe from the Gourmet Coolbook. I do hope you included the sambuca.
December 14, 2025 at 1:17 AM
My favorite cycling story about long_time managers, sponsors, and loyalty is 2010-2011 as JRB hunted new sponsors, team almost folded, until Europcar came along requiring Thomas Voeckler to stay. So Tommy turned down a big salary increase to save the team. Then wore yellow at the TdF for 11 days
November 21, 2025 at 12:45 PM
Sometimes you can manipulate the data frame to write the fomula as "Y ~ .". Then fit the model and extract the formula from the result.
November 17, 2025 at 12:19 PM
"We've always been at war with Eurasia."
November 15, 2025 at 11:12 AM
Or they steer healthier patients to trials A and sicker patients to trial B. Drug A succeeds; drug B fails. But they may be equivalent
November 11, 2025 at 12:11 PM
Imagine two trials. Same inclusion/exclusion. Drugs with different mechanisms. MDs steer patients to different trials based on information not formally assessed in the trials. I can imagine both trials getting better results than would hold in the general population defined by inc/exc.
November 11, 2025 at 12:09 PM
What caught my eye in your excerpt was the idea of physicians picking the trial patients. I noticed at both MDACC and OSU James that disease groups always had multiple trials running. Patients eligible for multiple trials weren't offered all of them. How does this affect RCT assumptions or analysis?
November 9, 2025 at 10:53 PM
Well, actually, those are LLMs, which are merely the latest impersonators of AI.
November 6, 2025 at 10:50 AM
D'you think they ever considered hyphenating their last names?
November 4, 2025 at 12:47 PM
I've been trying to keep the code for ny papers in GitLab projects for a long time. Probably since Keith Baggerly and I started pointing out the flaws in the Potti-Nevins series of papers from Duke in around 2008. How "historical" can papers be to be posted to the papers-with-code feed?
November 3, 2025 at 3:33 PM
A man after my own heart. I've been fighting this battle with trainees for decades. Also trying to convince them that active voice is better than passive. "Analyses were performed...".
November 3, 2025 at 11:49 AM
The chart shown already subtracted Pogi's points from UAE. They actually have 40,802 total for the year.
October 23, 2025 at 10:02 AM
Perhaps more importantly, if you took Pogačar's points away to create a new (virtual, one-man) team, then UAE would still be number one.
October 22, 2025 at 12:11 PM
Is there something magical about a 60-year anniversary?
October 20, 2025 at 1:20 AM
We need a yearly prize for most kilometers spent on the front of a chasing peloton. How many times would De Clerq have won it?
October 20, 2025 at 1:13 AM
Why do the French (and the Belgians) so frequently burden young riders with these early comparisons to Hinault (or Merckx)? How many future champions have fallen by the wayside because of the pressure? At least Isaac Del Toro only has to deal with being the next Pogaĉar.
October 20, 2025 at 1:10 AM
@antmccrossan.bsky.social called it out live. Lapeira and Gruel gave the opposite of a "master class", showing everyone how to give away the win of a bike race.
October 12, 2025 at 3:58 PM
Yes, but the Rubicon is considerably further south than that....
October 12, 2025 at 1:06 AM
I have found that it is a useful way to teach students how to understand p-values.
October 9, 2025 at 10:03 AM
As far as I can tell, the "vintage" leather cycling helmets meet all these requirements. Are those safer than the helmets being banned in non-time-trials?
October 8, 2025 at 11:02 AM
I'm still looking forward to Il Lombardia. But I wouldn't mind if some combination of teams figured out how to work together to make things harder for Pogaĉar. It's pretty clear that it will take a coalition.
October 8, 2025 at 12:20 AM
Will that convince Pogaĉar that it's time to win the Vuelta?
October 7, 2025 at 11:18 AM