A Small Publisher's Observations on Amazon and the Over-Production of AI Generated Content – aka “AI Slop”

  

A Small Publisher's Observations on Amazon and the Over-production of AI Generated Content – aka “AI Slop”

Over the last year or two, the phenomenon of “AI Slop” has increased significantly. For those few who haven’t yet been introduced to the term, it refers to content (text, code, pictures and video) that has been generated by generative AI algorithms. These are enormous deep-learning neural-net based programs that are fed tremendous amounts of data (called ‘training data’) that then generate new data, based on the input prompts given by human users.

There are a lot of issues with this process that generally fall into the categories of financial, ecological and social. There is a lot that can be said on any of these categories, but in this blog I am focusing on the social side, though that naturally has interactions with the financial and ecological sides of the question. My main issue will be about how generative AI is affecting human-created creative efforts and how the use of the technology is affecting the platforms (such as Amazon) that are used to transmit and market these creative efforts.

For what it is worth, here are some personal observations from running a small book business, primarily on Amazon:

• Over the last couple of years, the amount of our free promotion downloads has declined quite significantly, by about half.

• On the other hand, the amount of our paid sales (purchase and page-reads) has held fairly steady.

• New books seem to be harmed more than older books. This is partly due to the proliferation of content in general, but this has been super-accelerated by AI slop.

• One gets the impression that people are using publication date as a way to filter out AI generated content.

• If this is the case, Amazon and similar sites should make a policy of forcing people to identify AI usage when they upload. They already require this information when uploading a book, but they don’t seem to be revealing it to potential customers.

• This is probably a good example of how companies are conflicted – they want the increased content of AI generated content , but don’t want people to know which stuff is AI generated. In the long run, I don’t see how they can have it both ways.

On a related matter, I have noticed a lot of content on Quora that seems to be AI generated. I tend to avoid that and its prevalence on the site has made me less willing to contribute my own content and/or consume other content.

--------------------------------------------------------------------------------------------------

An article from the Globe and Mail (sourced as New York Times News Service) provides much of the data for the comments below. The title of the article is “Like many consumers, tech giants are sick of AI slop”.

The article goes on to describe the problems that companies such as Google, Meta (e.g. Facebook) and Amazon are having with generative AI content swamping their sites. These include things like google searches, Amazon’s book platform and Facebook’s feed. Naturally, there are plenty more companies facing the same problem (e.g. most other social media sites). It is, of course, ironic that the very companies that are spending extraordinary amounts of money and other resources on AI development are also being harmed by the output of these technologies. The phrase “hoist on their own petard” comes to mind.

Here are a few examples of what these companies have done:

• Spotify removed tens of millions of “spammy” uploads, presumably generated by AI or at least facilitated by AI in the upload process.

• Linkedin has added a button for users to notify them of content that seems like it was generated by AI.

• Youtube had wiped hundreds of thousands of channels in recent months, that appear to be low-quality AI generated ‘slop’.

• TikTok has removed hundreds of thousands of videos recently for violating their AI policies.

• Messaging apps, review sites, academic archives and dating services are also attempting to get rid of AI generated spam and slop.

• App sites (such as the Apple store) are being flooded with vibe-coded content, another type of AI generated content.

• X has banned apps that facilitate AI slop (it is estimated that over half of short tweets (less than 50 words) are AI generated).

• Substack is actively scanning posts for AI useage.

• Unfortunately, the article doesn’t say what Amazon is doing about slop, though it does note that it is an ongoing concern for them.

The article goes on to note that these companies have mixed and contrary strategies and incentives. On the one hand, AI generated content can initially increase engagement on their sites, just by sheer volume. That can mean gains in revenue and/or investment dollars. However, people can quickly tire of this deluge of AI slop and turn away for using the site. That can negate all those gains and ultimately poison the site.

This might all be considered part of the “tragedy of the commons”. Slop is cheap (sometimes free) to make and market. That attracts slop developers who see a quick buck, that is there for the taking. That works for a while, but the profusion of slop soon saturates the site and people then tend to desert it.

This problem may resolve itself, once the AI investment bubble is over. If Tech companies have to recover costs and make sustainable profits, the cost of using generative AI may become too high to be profitable for the slop producers. As an example, if it takes a lot of energy and computer chips to produce a novel, it may well be cheaper for novels to be written exclusively by humans.

There is also the fact that humans just naturally seem to dislike AI generated content, once the novelty wears off. This may also lead to the reduction, if not elimination, of AI slop.

--------------------------------------------------------------------------------------------------

And here is a science-fiction novel, featuring a galactic AI threat which might well be considered the ultimate in the AI threat (the book itself is human-generated).

The Magnetic Anomaly – A Science Fiction Novel

The novel version of The Magnetic Anomaly is now available on Amazon (ebook, print version, also audible version soon to come).

Summary

Two young Earth scientists, Alex and Mary, get caught up in a vast conflict, based on an accidental finding that they make at a diamond exploration camp in the far north. Eventually, this discovery leads them on a perilous quest and involvement in an interstellar war, a conflict of literally cosmic proportions with a rogue AI, which considers itself duty-bound to destroy sentient life (or, from its point of view, liberate sentient life).

During this time, they encounter a variety of enigmatic persons, as well as other entities, all of whom are also engaged in this struggle. With some of them they end up allying; with others, they contend in deadly conflict. The struggle takes them around the world, and eventually into the far reaches of the solar system.

During this time their relationship continues to evolve and deepen. Circumstances demand that they grow in courage and confidence, though within the scope of (mostly) normal people, as has been the case for countless people throughout history.

The story also showcases some historical events, relating to different versions of what went on in Antarctica, shortly after World War 2. Several versions of these events, official and unofficial UFO lore are given. This becomes a springboard to much of the action. Some philosophical and scientific concepts are also explored, though not in such a way as to get in the way of the plot.

Amazon:

U.S.: https://www.amazon.com/dp/B0F9J3HHFW

UK: https://www.amazon.co.uk/dp/B0F9J3HHFW

Canada: https://www.amazon.ca/dp/B0F9J3HHFW

France: https://www.amazon.fr/dp/B0F9J3HHFW

Italy: https://www.amazon.it/dp/B0F9J3HHFW

Netherlands: https://www.amazon.nl/dp/B0F9J3HHFW

Spain: https://www.amazon.es/dp/B0F9J3HHFW

Japan: https://www.amazon.co.jp/dp/B0F9J3HHFW

Mexico: https://www.amazon.com.mx/dp/B0F9J3HHFW

Australia: https://www.amazon.com.au/dp/B0F9J3HHFW

Germany: https://www.amazon.de/dp/B0F9J3HHFW

India:  https://www.amazon.in/dp/B0F9J3HHFW

Brazil: https://www.amazon.com.br/dp/B0F9J3HHFW

---------------------------------------------------------------------------------------------

does AI generated code suffer from the heartbreak of psoriasis

 does AI generated code suffer from the heartbreak of psoriasis?

While fooling around and experimenting with Gemini recently, I asked it the above question.  Below is its response.

By turns, I found it surprising, funny, and kind of sad.  It almost seemed like a plea for understanding from the LLM.  Having studied the underlying algorithms fairly intensely, I still believe that this is just a high-dimensional autocomplete program with a lot of post-training augmentation.  Nonetheless, its short-term resemblance to a human complaining about a health problem is quite fascinating.  I suppose a great deal of human communication is similar to autocomplete - i.e. we are in a version of autopilot quite a bit of the time.

I think its main points are valid, in terms of vibe-coding and AI generated code.  Organizations that lean on vibe-coding are probably storing up a lot of heartbreak in their futures.

-------------------------------------------------------------------------------------------------------------------------- 

Yes, AI-generated code suffers significantly from its own version of the "heartbreak of psoriasis"—an  irritating, persistent, and "scaly" accumulation of bugs and technical debt that can make it painful to maintain and deploy. [1]

While AI tools speed up development, they often introduce 1.7 times as many bugs as human developers. [1]

Here is how AI-generated code reflects this analogy:

• Scabs and Scaly Patches (Technical Debt): The code often looks clean on the surface but is packed with "cruft"—unnecessary, disorganized, or redundant code that behaves like flaky, inflamed skin on a project.

• Persistent Itch (Silent Logic Errors): A massive 60% of faults in AI code are "silent logic failures," which pass tests but fail in production. This causes ongoing, hard-to-diagnose irritation rather than a simple, easy-to-fix crash.

• Invisible Under-layers (Security Flaws): Up to 45% of AI-generated code contains security vulnerabilities. These hidden issues are deep-rooted, like psoriasis plaques that are not visible until they erupt as a breach.

• Chronic Condition (Lack of Understanding): Because AI doesn't truly understand context, it can produce code that looks good in isolation but acts like a "scar" when trying to integrate with existing, healthy code. [1, 2, 3, 4, 5]

The Bottom Line: Just as with "the heartbreak of psoriasis," you cannot simply ignore AI-generated code. It requires intense, consistent human oversight ("treatment") to stop the accumulation of errors from taking over the codebase. [1]

------------------------------------------------------------------------------------------------------------- 

 

 The Gambler’s Fallacy vs Regression to the Mean - Can a Bet be "Due to Win"

In statistics, betting, and life in general, we are often faced with the dilemma of what seem to be unusual streaks of bad or good luck. When these occur, it is natural to wonder when the streak will end, whether it will be soon or whether the wait will be interminable. It is natual to wonder, "Am I due for a change in luck?" From there, it is an easy transition to decide "I must be due".

Here is a story. This happened to me recently while betting the thoroughbreds at Santa Anita racetrack. To be precise, it was the horses that were at that California racetrack. I was at a local off-track racing venue of my hometown. But the illusion of ‘being there’ can be strong, especially when caught in the moment of a lucky or unlucky streak.

In my case it was the latter. I had bet 7 straight races, using a handicapping method that had worked surprisingly well during the previous two dozen forays to the track. It involved some fairly rigorous mathematical techniques and was well-grounded in the realities of horseracing (a speed handicapping method of my own invention). In all 7 races my horse had lost, with only a couple even being vaguely close. This was highly unusual, given the method’s previous success rate.

With the final race on deck, I had to dip into some cash reserves, which I hadn’t had to do for quite some time, as I had been playing with previous winnings up until then. I was tempted to skip the final race, but then I decided against it. You might just as well finish the job, I thought. I walked up to the betting machine to purchase a win ticket for the final race, with the grim forboding of a person stuck in a losing streak.

Then the thought occurred to me – "I must be due, there is no way this losing streak can continue. It seems so obvious. Plus, I have just bought a fresh betting voucher and have more than adequate financial resources anyway. So, why not step up the bet and win it all back?"

This, I proceeded to do.

The race was confusing. The colours that the horses were wearing were difficult to make out in the stretch drive. In fact, I made the notation in a small notebook: "Lost. Also ran. Back of field." So, it looked like I wasn’t due, after all.

But then the fellow next to me said "It looks like the 6 won."

"What’s that?" I thought. "That can’t be – I bet on the 6!"

But he was right. I had mistaken the colours of the 2 horse and the 6 horse. So I was due! Not only that, I had stepped up my bet by enough to cover all my previous losses, with some to spare.

So, was I due or was I not due? Or does that even make sense?

The Gambler’s Fallacy

In an introductory statistics course, you are likely to be introduced to the notion of the gambler’s fallacy. An example of this is coin flipping. If you flip a coin a fairly large number of times, and heads comes up more often than tails, it seems intuitively obvious that you can safely bet on tails and win the game (at least until things equal out).

Unfortunately, that is not a good strategy, as the coin doesn’t have a memory – it doesn’t know that tails are due. In the long run, the percentage of heads vs tails will tend to 50:50, but that can take a long time. Indeed, the preponderance of heads can continue to grow in raw numbers, even as the percentage of heads drops and approaches 50%. This does seem rather counter-intuitive, but it can be proved mathematically via the use of something called the mean value theorem.

Regression to the Mean

This is a different but related concept. Basically, it says that in some complex process (more complicated than coin-flipping), if a participant begins to perform much better (or worse) than its long-term average, the participant will regress back towards that average (fairly soon). So, for example, a hockey player on a heater (e.g. scoring streak) will likely go back to his long-term scoring, given enough time. Same for a horse (or some category of horse) at the race track. It is also works in genetics, where offspring tend to revert back to their ancestors in some characteristic, short of selective breeding.

 

How is this different from coin-tossing. I would venture to say that it is different because these processes do have a memory. Hockey players who go on a heater will generally be having some puck-luck, which can create as a high confidence level, at least temporarily. Opponents will become aware of the player’s scoring streak and actively try to subvert it (that’s what the game is all about). He will be shadowed by the opposing team’s best checker. He may be singled out for some rough play. The player’s confididence may then drop, as will his performance.

Similarly for horseraces. The horse has a memory, though it is difficult to say how much that affects a horse. However, trainers and jockeys definitely have memories that will affect the process. A hot horse may get better treatment from trainers and jockeys, thus helping its win streak. On the other hand, it may be run too often and get physically worn out. Its betting odds will drop, which may motivate trainers and jockeys to darken its form, so as to improve its odds. Horses running against it will have better odds, which may ensure that jockeys and trainers go to extra lengths to beat the hot horse.

So, basically, there are multiple participants in these types of complex games, all with their own goals and strategies (thousands to millions in the stock market). The combination of these actor’s strategies will tend to cancel each other out, thus tending to bring all the opponens back to long-term performance metrics.

Conclusion

So, which phenomenon prevails? The Gambler’s Fallacy (don’t base actions on the notion of something being ‘due’) or Regression to the Mean (you probably can base actions on something being due).

I don’t really know. I am conflicted. But I think it is probably a reasonable strategy for complex games (sports, stock markets, predictive markets) but not for relatively simple games.

-----------------------------------------------------------------------------------------------------------------------

And here are a couple of short works that you might enjoy, if these matters interest you.

A Dark Horse

In “A Dark Horse”, a gambler’s desire to hit a big win seems to lead him to make a Faustian bargain with a supernatural evil.  Or is it all just a string of unnaturally good luck?

The story is just $0.99 U.S. (equivalent in other currencies) and about 8000 words. It is also available on Kindle Unlimited and is occasionally on free promotion.

U.S.: https://www.amazon.com/dp/B01M9BS3Y5

U.K.: https://www.amazon.co.uk/dp/B01M9BS3Y5

Germany: https://www.amazon.de/dp/B01M9BS3Y5

France: https://www.amazon.fr/dp/B01M9BS3Y5

Italy: https://www.amazon.it/dp/B01M9BS3Y5

Netherlands: https://www.amazon.nl/dp/B01M9BS3Y5

Spain: https://www.amazon.es/dp/B01M9BS3Y5

Japan: https://www.amazon.co.jp/dp/B01M9BS3Y5

India: https://www.amazon.in/dp/B01M9BS3Y5

Mexico: https://www.amazon.com.mx/dp/B01M9BS3Y5

Brazil: https://www.amazon.com.br/dp/B01M9BS3Y5

Canada: https://www.amazon.ca/dp/B01MDMY2BR

Australia: https://www.amazon.com.au/dp/B01M9BS3Y5

Here’s an interesting review from Goodreads

(BTW, the writer has never met the reviewer and was not even aware of the review until very recently – You can look up the review on Goodreads, if you like)

A Dark Horse

Every gambler is bound to run out of luck eventually, right?

By far my favourite type of horror is psychological horror. I was quite pleased with how Mr. Olausen frightened his audience without spilling a single drop of blood or so much as hinting at anything gory. He knew exactly what hints to drop for us that made us deliciously dread the next scene simply by throwing out hints about who or what the dark horse might actually represent. This is the kind of stuff I love getting scared by, especially as Halloween approaches.

It would have been helpful to have more character development in this short story. While I certainly wouldn’t expect to see as much time spent on this as I would for a full-length novel, I did have trouble connecting to the main characters due to how little I knew about them and how much their personalities seemed to remain the same no matter what happened to them. If not for this issue, I would have felt comfortable choosing a much higher rating as the plot itself was well done.

I must admit to not knowing much about gambling at all, so I appreciated the brief explanations the narrator shared about how placing bets works and why some people have so much trouble walking away from a bet. While I will leave it up to experts on these topics to say how accurate everything was, I did enjoy learning more about the main character’s addiction and what he hoped to gain from betting on just one more game or race. It gave me a stronger sense of empathy for folks in his position.

A Dark Horse – A Gothic Tale was a deliciously chilling story for the Halloween season and beyond.

--------------------------------------------------------------------------------------------------------------------

And here is an account of a summer job on the railroad, during my undergraduate years.

One Summer Working on the Railroad

What follows is anaccount of a few weeks one long-ago summer, when I was 19 and was working for the Canadian National Railway (CNR) on a railroad construction gang, in the wilds of north-central British Columbia, Canada.

The journal is in the form of a letter, that was never sent. Decades later, I think it has an interesting historical resonance. At times I come off like a callow youth – I plead guilty as charged. I swore a lot more in those days than I do now, but in places the writing is surprisingly good, at least in my humble opinion. And the story has a compelling narrative arc.

There were a lot of interesting and dramatic events that occurred – a number of industrial accidents being the most serious. There were also some colorful characters on the crew, which resulted in some dramatic and at times amusing conflicts and altercations. I perhaps flatter myself by including myself in that number. Or perhaps I condemn myself – I’m not sure.

So, if you want to be reminded of one of those summer jobs that was kind of life-changing, read on. My story may just kick-start some memories of your own.

The memoir/journal is about 9,000 words, a length that can usually be read in an hour or so. It is priced at 99 cents U.S. (equivalent in other currencies) and is free on Kindle Unlimited. Periodically, it will be offered as a free promotion.

U.S.: https://www.amazon.com/dp/B0CN661P8Z

UK: https://www.amazon.co.uk/dp/B0CN661P8Z

Canada: https://www.amazon.ca/dp/B0CN661P8Z

Australia: https://www.amazon.com.au/dp/B0CN661P8Z

India: https://www.amazon.in/dp/B0CN661P8Z

Germany: https://www.amazon.de/dp/B0CN661P8Z

France: https://www.amazon.fr/dp/B0CN661P8Z

Spain: https://www.amazon.es/dp/B0CN661P8Z

Italy: https://www.amazon.it/dp/B0CN661P8Z

Netherlands: https://www.amazon.nl/dp/B0CN661P8Z

Japan: https://www.amazon.co.jp/dp/B0CN661P8Z

Brazil: https://www.amazon.com.br/dp/B0CN661P8Z

Mexico: https://www.amazon.com.mx/dp/B0CN661P8Z

Pi Day 2026 –PI and the Horse Races

 

Pi Day 2026 –PI and the Horse Races

Since I have been playing around with picking the ponies lately, I thought I might try to look into any PI Day angles for that. So, I settled on examining how the circular nature of a racetrack affects the horse-races, in this case the harness races. However, the general sense of this would apply to any sort of race run around a track, including thoroughbred horses.

 

The above is a plan view of a simplified racetrack, comprised of two circles connected by straight lines, such that the lines intesect the circles, tangent to the radii (more or less, given the limitations of a spreadsheet construction).

The blue inner radii and connecting line represents the path of a horse that runs inside, along the rail. It is meant to be one-mile, the distance of many harness-racetracks. The green outer radii and connecting line represents the path of a horse that runs on an path to the outside of the first horse. The distance between the two paths is exagerated.

As the diagram shows, it is obvious that the outer horse will run a greater distance than the inner horse, which will be related to the difference in the radii of the inner and outer circles.

Since the simplified track is composed partly of circles, PI is bound to come into the calculations. That is shown in the cut/paste of some calculations from a spreadsheet. The main result is that the outer horse runs about 10 yards farther than the inner horse, if is it parked outside the inner horse on the turns. The phrase "parked out" just means that it is running outside of the innermost horse and remains "parked" in the position for the bulk of the race. That is the case when it is parked out by one horse. If it is on the outside of two horses, the extra distance will be nearly 21 yards, and so on. Note that this extra distance is on the turns; on the straight sections, both horses run the same distance.

The next set of calculations related to times and velocities. If the inside horse finishes the one-mile course in 2 minutes (a typical standard for harness horses), that implies a velocity of 14.67 yards per second (that’s 30 miles per hour). Were the horse that is parked one wide to go that same velocity, it would take 120.71 seconds, so a difference of about 7 seconds.

 

A typical length for a horse and sulky is about 12 to 14 feet, or about 4 yards. At about 15 yards per second, a horse would cover nearly 4 seconds in the time it takes to travel that distance. So, that would imply that the outside horse would finish the race about 2.6 lengths behind the winner.

In order to finish in a dead heat, the outer horse would have to step up its average velocity from 14.67 yards per second to 14.75 yard per second. That doesn’t sound like much, but over 2 minutes it adds up. For a horse that was parked outside for more widths than this, the numbers would scale appropriately.

So, where position the horse and when to challenge for the lead is significantly impacted by the extra distance to travel, due to the differing radii in the corners for inner and outer horses. And since corners are partial circles, PI comes into play.

This is one of the main factors that a driver must consider. Of course there are many other nuances. Some horses just naturally want to run in the lead, while others prefer to close at the end of the race. There can be advantages to running second, just behind the leader, in a similar way that this occurs in bicycle races ("drafting"). This can save a lot of energy, for that final sprint to the finishing line. But there is also a chance of being "boxed in" (stuck in traffic), and being prevented from making that move.

So, there is a lot of math in this, including PI, so I will go to the harness races on March 14 and have some PI.

-----------------------------------------------------------------------

So, now that you have dutifully read some Pi Day math, you should read a science fiction book, or even better, a whole series. Book 1 of the Witches’ Stones series even includes a reference to pi.:

The Witches’ Stones

Or, you might prefer, the trilogy of the Witches’ Stones (they’re psychic aliens, not actual witches), which follows the interactions of a future Earth confederation, an opposing galactic power, and the Witches of Kordea. It features Sarah Mackenzie, another feisty young Earth woman).

https://www.amazon.com/dp/B008PNIRP4

https://www.amazon.co.uk/dp/B008PNIRP4

 

 

The Magnetic Anomaly: A Science Fiction Story

“A geophysical crew went into the Canadian north. There were some regrettable accidents among a few ex-military who had become geophysical contractors after their service in the forces. A young man and young woman went temporarily mad from the stress of seeing that. They imagined things, terrible things. But both are known to have vivid imaginations; we have childhood records to verify that. It was all very sad. That’s the official story.”

A short story of about 6000 works.

https://www.amazon.com/dp/B0176H22B4

https://www.amazon.co.uk/dp/B0176H22B4

 

 

The Zoo Hypothesis or The News of the World: A Science Fiction Story

In the field known as Astrobiology, there is a research program called SETI, The Search for Extraterrestrial Intelligence. At the heart of SETI, there is a mystery known as The Great Silence, or The Fermi Paradox, named after the famous physicist Enrico Fermi. Essentially, he asked “If they exist, where are they?”.


Some quite cogent arguments maintain that if there was extraterrestrial intelligence, they should have visited the Earth by now. This story, a bit tongue in cheek, gives a fictional account of one explanation for The Great Silence, known as The Zoo Hypothesis. Are we a protected species, in a Cosmic Zoo? If so, how did this come about? Read on, for one possible solution to The Fermi Paradox.

The short story is about 6300 words, or about half an hour at typical reading speeds.

https://www.amazon.com/dp/B076RR1PGD

https://www.amazon.co.uk/dp/B076RR1PGD

Grok’s mistaken report on X for Oilers-Sens game, March 3, 2026

Grok’s Version

Bouchard's OT Winner Caps Oilers' Wild Comeback Over Senators

Last updated 1 hour ago

Evan Bouchard fired the power-play winner at 1:50 of OT after Brady Tkachuk hooked Connor McDavid, securing Edmonton's 5-4 comeback victory at Rogers Place. Draisaitl scored twice and added three assists in a five-point night, while McDavid notched two helpers to hit 105 points; Zach Hyman tied it with an empty-netter. Bouchard, the first defenseman to 70 points this season, extended his hot streak with seven goals and 15 assists in 11 games, drawing praise from Draisaitl as elite company with Makar and Hughes. The win lifts Edmonton to 30-24-8, two points back in the Pacific with games in hand.

This story is a summary of posts on X and may evolve over time. Grok can make mistakes, verify its outputs.

 

Official Oiler Website version

Hyman scored with the net empty late in regulation & Bouchard completed the comeback 1:50 into OT for the #Oilers to cap off a 5-4 victory over the Senators.

 

Comparison and Analysis

Grok's report said "Zach Hyman tied it with an empty-netter", which made no sense. I think it took "scored with the net empty late in regulation" and "tied it with an empty netter" to mean the same thing. Shows that there is no real understanding in LLMs, just stringing together likely word correlations, based on input data.

Also, the Oilers do not have "games in hand". In fact, the other teams in the division have played from 1 to 3 fewer games than the Oilers, at the time of this Grok summary (first column below).

I just like to point these things out, occasionally.  It is worth keeping this in mind, when reading LLM summaries of serious news, though.

Updated: Goaltending Since the Oilers-Penguins Trade

Now that the Olympics are done, it is a good time to review how the Oilers-Penguins trade worked out, using a longer time-line than the previous blog.

Tristan Jarry Update

For reference, the last blog was when Jarry was at Game 20. At that time, things were looking pretty good for him. Since then, the graph has turned way down, including a couple of .750 games, though it has not all been bad (one .900 game). This up-and-down performance was also seen during his time with Pittsburgh, though his record was somewhat better there, to this point in time. That may be due to the Penguins playing a better defensive game than the Oilers, or it may just be the random variation that is not unusual in such data. Consistency, in any high level activity, is a lot harder than the average person thinks.

 

Overall, since the trade, Tristan has had 1 excellent game (a shutout), 4 very good games (around .900), 4 so-so games (around .850) and 2 stinkers (around .750).

Jarry is still above water since the trade, in terms of (actual points/possible points) = 64%. However, that’s a drop from 75%, which was his record since the trade in the original post. Interestingly, the 64% is exactly the same as his record with the Penguins, earlier in the season.

Stuart Skinner Update

Next, we will look at Stuart Skinner’s updated performance since the trade, since the trade was Jarry for Skinner. For reference, Stu’s last game before this graph was updated was game 32).

 The 4 games since the original graph are similar to Jarry’s, in terms of the up-and-down nature of the data. He had two excellent games (around .950) and two howlers (both under .800). So, it could be argued that the teams traded one inconsistent goalie for a different inconsistent goalie.

Overall, since the trade, Stu has had 5 really good games (around .950), 2 pretty good games (around .900), 3 so-so games (around .850) and 3 howlers (below .800).

Skinner’s (actual points/possible points) = 66% since the trade, whereas it had been 57% in the original post. So, a decent improvement. For comparison, his record with the Oilers was 52%.

Conner Ingram Update

Next, we will look at Conner Ingram. Though he wasn’t part of the trade, the trade certainly affected his career path. For reference, his last game before the update of the graph was game 9, the shutout.

 

Since that time, Ingram has not done very well, either. Two games were semi-respectable and 2 were bad, below .800.

 

Calvin Pickard Update

Pickard has not played any Oilers games since the trade, though he was with the team for some of that time. However he was ultimately put on waivers, passed waivers, then sent down to the Oilers AHL team in Bakersfield, California. He hasn’t played since being sent down, as far as I know. It turns out the the farm team has a couple of very good goalie prospects, so it may be difficult to slot him in.

It is too bad, because he did make some really stand-out performances in the playoffs, during the previous couple of years, bailing out the team when Skinnter faltered. He is also known for being a good team player (‘good in the room’) and an overall nice guy. However, he is well into his 30s and has had a long and interesting career (it would make a good hockey book). Plus, even with relatively low pay for a hockey player, he had make a lot more money than 99% of the population did over that time. So there’s that.

-------------------------------------------------------------------------------------------------------------------------

Original Post

The Edmonton Oilers hockey club has long been seen as having inconsistent goaltending – sometimes their goalies have been stellar, other times seemingly hopeless. The latest tandem to exhibit that tendency was Stuart Skinner and Calvin Pickard. Many observers felt that more reliable goaltending would have led to at least one Stanley Cup victory, rather than a 7 and 6 game losses, as was the case in 2023-24 and 2024-25.

Thus, in mid-December 2025 they made a key goalie trade with the Pittsburgh Penguins, swapping Stuart Skinner for Tristan Jarry. As a side-effect of that trade, they also called up Conner Ingram from their AHL affiliate, since Jarry had experienced an injury shortly after the trade. The net effect (pun intended) was that they transitioned from Skinner-Pickard to Jarry-Ingram, with Calvin Pickard in a third goalie role, at least for a while, as insurance should one of the other two become injured again.

At the time of this blog, it has been a little over a month since the trade. So, how has it gone? Read below for some statistical evidence.

NOTE, next day: Hilariously enough, Oilers and Penguins played a game the very night that I published this blog and Jarry absolutely bombed (0.727 in that game), sending his save pct to 0.881 (from 0.904) in saves/shots and 0.878 (from 0.903) in average of game save percentages. His win pct dropped to 57% (from 67%), and his point pct went down to 64% (from 75%).

As I pointed out in the blog: his trend with the Oilers has only a limited number of games, so it should be taken with a large measure of caution.

Statisticians say this all the time, because it’s true.

 

Tristan Jarry

Jarry has played 6 games, as of Dec 21, 2026 for the Oilers. Previous to that, he had played 14 games with the Penguins. His Oilers save percentage since the trade has been over 0.900 , customarily thought of as an important benchmark. In terms of overall saves/shots, it is 0.904, while in terms of averaging the save percentage in each game, it is at 0.903. Note that the two methods can yield slightly different results, depending on how many outlier games the goalie has had (i.e. very much better or worse than usual).

His record with the Penguins this year, was not all that different from that with the Oilers, in terms of save percentage, at 0.909 for saves/shots and 0.895 when averaging the save percentages in each game. This result shows how the two methods can give different answers. That is further demonstrated by the high ups and downs of his Penguins games, as shown in the graph.

The graph shows that Jarry’s linear trend with the Penguins was slightly downward, while the trend with the Oilers it has been upward. That said, his trend with the Oilers has only a limited number of games, so it should be taken with a large measure of caution.

Jarry won 4 out of these 6 Oiler games, for a win percentage of 67%. In terms of points, he has played net for 9 of 12 potential points, for a points average of 75%.

For the Penguins earlier in the year, his win percentage in 14 games was 64%, with his points percentage coming in at 68%.

So, as far as Jarry’s play goes, the trade has worked out well so far for him and for the Oilers.

Stuart Skinner

Stuart Skinner has played 9 games with the Penguins, after having played 23 games with the Oilers previous to that. His Penguins save percentage since the trade has been a bit under 0.900. In terms of overall saves/shots, it is 0.891, while in terms of averaging the save percentage in each game, it is at 0.895. Note again that the two methods can yield different results, depending on how many outlier games the goalie has had (i.e. very much better or worse than usual).

The graph shows that Skinner’s linear trend with the Oilers was downward, while with the Penguins it has been upward. That said, his trend with the Penguins has only a limited number of games, so it should also be taken with a large measure of caution.

His record with the Oilers this year, was not as good as it has been with the Penguins. His Oilers save percentage was 0.891 for saves/shots and 0.872 when averaging the save percentages in each game. Again, this result shows how the two methods can give quite different answers. That is further demonstrated by the high ups and downs of his Oilers games, as shown in the graph. It is also the case that his games with the Penguins have had a fair bit of variance.

Skinner won 5 out of the 9 Penguin games, for a win percentage of 56%. In terms of points, he has played net for 10 of 18, also for a points average of 56%.

For the Oilers earlier in the year, his win percentage in 23 games was 48%, with his points percentage coming in at 57%.

So, as far as Skinner’s play goes, the trade has worked out well so far for him (higher save percentage), and it has probably been a slight loss for the Penguins. However, it is a small sample size, so that may change as the season progresses.

Conner Ingram

Conner Ingram was not directly involved in the trade, though it is probable that the trade did make his call-up from the AHL affiliate necessary. Since then, he has answered the bell quite decisively, with a 0.917 save percentage for the total saves/shots measure, and 0.920 for the average of averages measure.

The graph shows that Ingram’s linear trend with the Oilers has been upward. That said, it is a limited number of games (9 in all), so it should also be taken with a measure of caution.

Ingram has won 5 out of 9 starts, for a win percentage of 56%. In terms of points, he has played for 11 points out of a possible 18, for a points percentage of 61%.

So, as far as Ingram’s play goes, the trade has worked out well so far for him. It has resulted in him being called up to the NHL team, where he has posted an impressive save percentage as well as a very good win and points percentage. However, it is a small sample size, so that may change as the season progresses.

Calvin Pickard

Calvin Pickard was not part of the trade, but it has also had an impact on him. He had several starts while Jarry was injured, more than he otherwise would have had. Previous to the trade his save percentage was mediocre at best, with 0.851 in total saves/shots and 0.848 in average of game save percentages. Since the trade those figures have improved considerably, to 0.901 for the former measure and 0.896 for the latter.

The graph shows that Pickard’s linear trend with the Oilers has been upward, largely due to the good play since the trade. That said, his post-trade sample of games (6 in all) is relatively small, so it should be taken with a measure of caution.

 

Pickard won 3 out of the 6 post-trade games, for a win percentage of 50%. In terms of points, he has played net for 6 of 12, also for a points average of 50%.

For the Oilers earlier in the year, his win percentage in 10 games was 30%, with his points percentage coming in at 40%.

So, as far as Pickard’s play goes, the trade has worked out fairly well so far for him. His save percentages, win percentages and points percentages have improved. However, it is a small sample size, so that may change as the season progresses.

It is not clear how long the Oilers can afford to carry three goalies. Pickard’s improved performances have made that a tougher call. He could be picked up by some other team, should they choose to send him to the AHL affilate. Arguably, his post-trade play has made that a greater possibility than it had been earlier in the year.

Overall Result

The data support the idea that the Oilers have indeed improved their goaltending situation by the trade.

As for the Penguins, they may have taken a slight hit from the trade.

However, it is still early days, so those results could change.

--------------------------------------------------------------------------------------------------------------------

A Dark Horse

In “A Dark Horse”, a gambler’s desire to hit a big win seems to lead him to make a Faustian bargain with a supernatural evil.  Or is it all just a string of unnaturally good luck?

The story is just $0.99 U.S. (equivalent in other currencies) and about 8000 words. It is also available on Kindle Unlimited and is occasionally on free promotion.

U.S.: https://www.amazon.com/dp/B01M9BS3Y5

U.K.: https://www.amazon.co.uk/dp/B01M9BS3Y5

Germany: https://www.amazon.de/dp/B01M9BS3Y5

France: https://www.amazon.fr/dp/B01M9BS3Y5

Italy: https://www.amazon.it/dp/B01M9BS3Y5

Netherlands: https://www.amazon.nl/dp/B01M9BS3Y5

Spain: https://www.amazon.es/dp/B01M9BS3Y5

Japan: https://www.amazon.co.jp/dp/B01M9BS3Y5

India: https://www.amazon.in/dp/B01M9BS3Y5

Mexico: https://www.amazon.com.mx/dp/B01M9BS3Y5

Brazil: https://www.amazon.com.br/dp/B01M9BS3Y5

Canada: https://www.amazon.ca/dp/B01MDMY2BR

Australia: https://www.amazon.com.au/dp/B01M9BS3Y5

Here’s an interesting review from Goodreads

(BTW, the writer has never met the reviewer and was not even aware of the review until very recently – You can look up the review on Goodreads, if you like)

A Dark Horse

Every gambler is bound to run out of luck eventually, right?

By far my favourite type of horror is psychological horror. I was quite pleased with how Mr. Olausen frightened his audience without spilling a single drop of blood or so much as hinting at anything gory. He knew exactly what hints to drop for us that made us deliciously dread the next scene simply by throwing out hints about who or what the dark horse might actually represent. This is the kind of stuff I love getting scared by, especially as Halloween approaches.

It would have been helpful to have more character development in this short story. While I certainly wouldn’t expect to see as much time spent on this as I would for a full-length novel, I did have trouble connecting to the main characters due to how little I knew about them and how much their personalities seemed to remain the same no matter what happened to them. If not for this issue, I would have felt comfortable choosing a much higher rating as the plot itself was well done.

I must admit to not knowing much about gambling at all, so I appreciated the brief explanations the narrator shared about how placing bets works and why some people have so much trouble walking away from a bet. While I will leave it up to experts on these topics to say how accurate everything was, I did enjoy learning more about the main character’s addiction and what he hoped to gain from betting on just one more game or race. It gave me a stronger sense of empathy for folks in his position.

A Dark Horse – A Gothic Tale was a deliciously chilling story for the Halloween season and beyond.

--------------------------------------------------------------------------------------------------------------------

The Magnetic Anomaly – A Science Fiction Novel

The novel version of The Magnetic Anomaly is now available on Amazon (ebook, print version, also audible version soon to come). 

Summary

Two young Earth scientists, Alex and Mary, get caught up in a vast conflict, based on an accidental finding that they make at a diamond exploration camp in the far north. Eventually, this discovery leads them on a perilous quest and involvement in an interstellar war, a conflict of literally cosmic proportions with a rogue AI, which considers itself duty-bound to destroy sentient life (or, from its point of view, liberate sentient life).

During this time, they encounter a variety of enigmatic persons, as well as other entities, all of whom are also engaged in this struggle. With some of them they end up allying; with others, they contend in deadlyconflict. The struggle takes them around the world, and eventually into the far reaches of the solar system.

During this time their relationship continues to evolve and deepen. Circumstances demand that they grow in courage and confidence, though within the scope of (mostly) normal people, as has been the case for countless people throughout history.

The story also showcases some historical events, relating to different versions of what went on in Antarctica, shortly after World War 2. Several versions of these events, official and unofficial UFO lore are given. This becomes a springboard to much of the action. Some philosophical and scientific concepts are also explored, though not in such a way as to get in the way of the plot.

Amazon:

U.S.: https://www.amazon.com/dp/B0F9J3HHFW

UK: https://www.amazon.co.uk/dp/B0F9J3HHFW

Canada: https://www.amazon.ca/dp/B0F9J3HHFW

France: https://www.amazon.fr/dp/B0F9J3HHFW

Italy: https://www.amazon.it/dp/B0F9J3HHFW

Netherlands: https://www.amazon.nl/dp/B0F9J3HHFW

Spain: https://www.amazon.es/dp/B0F9J3HHFW

Japan: https://www.amazon.co.jp/dp/B0F9J3HHFW

Mexico: https://www.amazon.com.mx/dp/B0F9J3HHFW

Australia: https://www.amazon.com.au/dp/B0F9J3HHFW

Germany: https://www.amazon.de/dp/B0F9J3HHFW

India:  https://www.amazon.in/dp/B0F9J3HHFW

Brazil: https://www.amazon.com.br/dp/B0F9J3HHFW

---------------------------------------------------------------------------------------------