Philipp Wacker
@phkwacker.bsky.social
Senior Lecturer in Statistics, University of Canterbury (NZ).
he/him
Bayesian stats, Applied Math, filtering & sampling
pls send me awesome vegan waffle recipes!
he/him
Bayesian stats, Applied Math, filtering & sampling
pls send me awesome vegan waffle recipes!
A surprisingly fun game to play with your kids on long car trips: Count up, but only numbers that you see on the way. A 1 on a number plate. A 2 on a phone number on an ad, a 3 on a bus stop. Gets trickier once you're in double digits and beyond.
February 10, 2026 at 5:42 AM
A surprisingly fun game to play with your kids on long car trips: Count up, but only numbers that you see on the way. A 1 on a number plate. A 2 on a phone number on an ad, a 3 on a bus stop. Gets trickier once you're in double digits and beyond.
When I have to use a journal's LaTeX template for a paper slowly chip down the number of new compile errors from about 500 down to 0 (hopefully!), I feel like one of these lumberjacks carving out a grizzly bear with a chainsaw.
February 2, 2026 at 9:47 AM
When I have to use a journal's LaTeX template for a paper slowly chip down the number of new compile errors from about 500 down to 0 (hopefully!), I feel like one of these lumberjacks carving out a grizzly bear with a chainsaw.
I'm also a big stan of Monte Carlo methods tbh
January 30, 2026 at 8:15 AM
I'm also a big stan of Monte Carlo methods tbh
team works together and solves the case. There is also the gruff-but-actually-nice police chief (Vimes/Bogo), a crime syndicate of frost-loving giants (trolls/polar bears). Ah, and Ankh-Morpork's ruler is Lord VETinari?
January 29, 2026 at 9:00 AM
team works together and solves the case. There is also the gruff-but-actually-nice police chief (Vimes/Bogo), a crime syndicate of frost-loving giants (trolls/polar bears). Ah, and Ankh-Morpork's ruler is Lord VETinari?
police force in the metropolis (Ankh-Morpork/Zootopia). Soon it becomes apparent there is brooding conflict between the factions living in the city (Humans, Dwarves, Undead/Predator and Prey animals). An unknown bad guy tries to turn these factions against each other. But the mixed-faction police
January 29, 2026 at 9:00 AM
police force in the metropolis (Ankh-Morpork/Zootopia). Soon it becomes apparent there is brooding conflict between the factions living in the city (Humans, Dwarves, Undead/Predator and Prey animals). An unknown bad guy tries to turn these factions against each other. But the mixed-faction police
After rewatching Zootopia I'm now convinced it should be considered an adaptation of Terry Pratchett's Nightwatch canon. The plot of both is like this: A young and naive person from the country (Carrot Ironfoundersson in NW, a literal bunny with a carrot farming background in Zootopia) joins the
January 29, 2026 at 9:00 AM
After rewatching Zootopia I'm now convinced it should be considered an adaptation of Terry Pratchett's Nightwatch canon. The plot of both is like this: A young and naive person from the country (Carrot Ironfoundersson in NW, a literal bunny with a carrot farming background in Zootopia) joins the
This is hilarious: arxiv.org/abs/2601.20123
Optimal illness policy for an unethical daycare center
While businesses are typically more profitable if their workers and communities are minimally exposed to diseases, the same is not true for daycare centers. Here it is shown that a daycare center coul...
arxiv.org
January 29, 2026 at 6:51 AM
This is hilarious: arxiv.org/abs/2601.20123
Any resources for how to make oral assessments better in terms of equity (in particular, minimising impact of the assessor's implicit biases)?
January 26, 2026 at 7:29 AM
Any resources for how to make oral assessments better in terms of equity (in particular, minimising impact of the assessor's implicit biases)?
might be just manageable.
3. Give them a speech about how bad GenAI is for their learning if they make it do their homework and otherwise keep everything as is (and hope for the best).
4. ??
3. Give them a speech about how bad GenAI is for their learning if they make it do their homework and otherwise keep everything as is (and hope for the best).
4. ??
January 26, 2026 at 1:07 AM
might be just manageable.
3. Give them a speech about how bad GenAI is for their learning if they make it do their homework and otherwise keep everything as is (and hope for the best).
4. ??
3. Give them a speech about how bad GenAI is for their learning if they make it do their homework and otherwise keep everything as is (and hope for the best).
4. ??
1. Get rid of uninvigilated assessment altogether (sad)
2. Make it possible to pass the course without doing any assignments, but if you want an A, you have to do an extra assignment (and I will ask you questions about it in my office). If enough students are OK with "just" getting a B, that
2. Make it possible to pass the course without doing any assignments, but if you want an A, you have to do an extra assignment (and I will ask you questions about it in my office). If enough students are OK with "just" getting a B, that
January 26, 2026 at 1:07 AM
1. Get rid of uninvigilated assessment altogether (sad)
2. Make it possible to pass the course without doing any assignments, but if you want an A, you have to do an extra assignment (and I will ask you questions about it in my office). If enough students are OK with "just" getting a B, that
2. Make it possible to pass the course without doing any assignments, but if you want an A, you have to do an extra assignment (and I will ask you questions about it in my office). If enough students are OK with "just" getting a B, that
I want to redesign the assessment of my 2nd year probability and random processes course in a way that is more robust against GenAI. So far I just had assignments with exercises that can now trivially be solved by LLMs out there. Any ideas? So far I'm thinking:
January 26, 2026 at 1:07 AM
I want to redesign the assessment of my 2nd year probability and random processes course in a way that is more robust against GenAI. So far I just had assignments with exercises that can now trivially be solved by LLMs out there. Any ideas? So far I'm thinking:
One of my favourite things about my job as a dad: Pretending to be annoyed when the kids sneakily steal and munch all the cut up cucumbers and tomatoes that I "need for other things"
January 18, 2026 at 5:41 AM
One of my favourite things about my job as a dad: Pretending to be annoyed when the kids sneakily steal and munch all the cut up cucumbers and tomatoes that I "need for other things"
Try listening to a lecture about it, I always get terrible Gibbs ringing in my ears.
January 18, 2026 at 12:15 AM
Try listening to a lecture about it, I always get terrible Gibbs ringing in my ears.
Ooh, also related to Bayesian updates (the || operation is how covariance matrices are "added")
January 7, 2026 at 12:42 AM
Ooh, also related to Bayesian updates (the || operation is how covariance matrices are "added")
I assume garments are sorted according to their number of holes
December 21, 2025 at 7:20 AM
I assume garments are sorted according to their number of holes
Stipend NZD 35k/yr (paid to student) + full PhD fees covered.
Start Apr 2026 (flexible).
Apply by 31 Jan 2026 → [email protected]
Full ad: tinyurl.com/td79d687
Please share!
Start Apr 2026 (flexible).
Apply by 31 Jan 2026 → [email protected]
Full ad: tinyurl.com/td79d687
Please share!
Funded PhD Position: Bayesian Optimal Experimental Design
tinyurl.com
December 15, 2025 at 9:55 PM
Stipend NZD 35k/yr (paid to student) + full PhD fees covered.
Start Apr 2026 (flexible).
Apply by 31 Jan 2026 → [email protected]
Full ad: tinyurl.com/td79d687
Please share!
Start Apr 2026 (flexible).
Apply by 31 Jan 2026 → [email protected]
Full ad: tinyurl.com/td79d687
Please share!
Funded PhD (3 yrs): Bayesian Optimal Experimental Design
Univ. of Canterbury, Christchurch, New Zealand 🇳🇿 | Marsden Fund
Math & computation for Bayesian OED in inverse problems (optimal control, filtering).
Women & others underrepresented in maths/stats in particular encouraged to apply.
Univ. of Canterbury, Christchurch, New Zealand 🇳🇿 | Marsden Fund
Math & computation for Bayesian OED in inverse problems (optimal control, filtering).
Women & others underrepresented in maths/stats in particular encouraged to apply.
Funded PhD Position: Bayesian Optimal Experimental Design
tinyurl.com
December 15, 2025 at 9:55 PM
Funded PhD (3 yrs): Bayesian Optimal Experimental Design
Univ. of Canterbury, Christchurch, New Zealand 🇳🇿 | Marsden Fund
Math & computation for Bayesian OED in inverse problems (optimal control, filtering).
Women & others underrepresented in maths/stats in particular encouraged to apply.
Univ. of Canterbury, Christchurch, New Zealand 🇳🇿 | Marsden Fund
Math & computation for Bayesian OED in inverse problems (optimal control, filtering).
Women & others underrepresented in maths/stats in particular encouraged to apply.
I think we need another symbol for the product integral, like this:
December 7, 2025 at 10:19 PM
I think we need another symbol for the product integral, like this:
Am I the only one who is annoyed that we don't have a special symbol for the product integral? The \Sigma becomes and \int when we take the limit in the (Riemann) integral approximation, but the \Pi just stays a \Pi when we make a product integral!
December 7, 2025 at 10:19 PM
Am I the only one who is annoyed that we don't have a special symbol for the product integral? The \Sigma becomes and \int when we take the limit in the (Riemann) integral approximation, but the \Pi just stays a \Pi when we make a product integral!
I mean, still looks cool, though?
November 8, 2025 at 3:22 AM
I mean, still looks cool, though?
At the same time, unfortunately, every 5th number is evil arxiv.org/abs/2510.03940
Every Fifth Real Number is Evil
Fifteen years ago, then-Carleton-undergrad Isaac Hodes, proved that the Golden Ratio is evil. In this modest contribution to human knowledge, we show that in fact, every fifth real number is evil, and...
arxiv.org
October 20, 2025 at 7:42 PM
At the same time, unfortunately, every 5th number is evil arxiv.org/abs/2510.03940
This is brutal. Without even mentioning all the science funding cuts youtu.be/h-yjkAXxnMA?...
Honest Government Ad | Visit New Zealand!
YouTube video by thejuicemedia
youtu.be
October 14, 2025 at 8:24 PM
This is brutal. Without even mentioning all the science funding cuts youtu.be/h-yjkAXxnMA?...
I am sure he will take it well when he won't have won the Nobel Peace Prize in four hours.
October 10, 2025 at 7:01 AM
I am sure he will take it well when he won't have won the Nobel Peace Prize in four hours.
I personally think about it like this: Is learning this going to be "good" (enjoyable, pay-off in the future, retention realistic)? If so, I do it myself. Hacking alignments in LaTeX, for me, does not fall in this category and I would also use LLMs for that.
October 8, 2025 at 11:18 PM
I personally think about it like this: Is learning this going to be "good" (enjoyable, pay-off in the future, retention realistic)? If so, I do it myself. Hacking alignments in LaTeX, for me, does not fall in this category and I would also use LLMs for that.
But I feel you, it is eggshausting when that happens.
October 7, 2025 at 6:43 AM
But I feel you, it is eggshausting when that happens.