Rendered at 13:03:05 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
wespiser_2018 19 hours ago [-]
Author here: I did a quick experiment with my Greyhound, Bebop, to figure out the treat he prefers best using pair-wise comparison analyzed with the Bradley-Terry model. Same tech as Elo scores in chess, and several other places! Enjoy!
zthrowaway 3 hours ago [-]
Really weird question but what boots are you wearing in your blog photo? They look wide enough for my weird feet. Thanks
gwern 14 hours ago [-]
One thing I'm not following is how the side/order bias is being handled. OP measures a IMO very large bias towards left-hand treats, but it is unclear how that is handled (ahem)? Skimming https://github.com/adamwespiser/best-dog-treat/blame/main/an... doesn't help me understand if it is being modeled as a covariate to adjust for the bias or if it was dealt with by construction (eg. by always offering pairs twice, swapping hands), or what?
Incidentally OP if you want to make it more adaptive, you can just fit the B-T model each time, and grab a posterior sample of what the best pair is, and test that, which turns out to be Thompson sampling. I did this for fun with blind taste-testing of mineral waters: https://gwern.net/water
markerz 9 hours ago [-]
I suspect it’s by construction, looks like they’re offering every permutation of treat choice twice.
Below, I filtered for A vs E, the top two choices. Notice how they switch left and right hand each time:
A/E :: A
E/A :: A
A/E :: E
E/A :: E
A/E :: E
wespiser_2018 34 minutes ago [-]
Author here, yes, that's correct.
The initial set-up was to do 3 comparisons of the 5 treats (30 trials), alternate between right and left hand, then write a quick python script to randomize the order.
A bit more than halfway through the experiment, I ran the model and realized that A/D/E were the only contenders left, so I removed the B/C trials and added more A/D/E trials.
wespiser_2018 13 hours ago [-]
That's awesome! Thank you!
buildsjets 18 hours ago [-]
The best dog treat is always the treat the OTHER dog is eating.
ses1984 11 hours ago [-]
My one dog hates broccoli but she will choke it down just so the other one can’t have it.
savethejoy 11 hours ago [-]
Oh my goodness.
weberer 10 hours ago [-]
Or cat turds
buildsjets 9 hours ago [-]
Kitty roca confirmed. Our Husky was a goon for it. At least it forced us to keep the boxes clean several times a day.
NordStreamYacht 9 hours ago [-]
Good grief.
Why would you give your dog those as treats?
buildsjets 3 hours ago [-]
You don’t need to give it to them, they just graze at the conveniently provided buffet.
NordStreamYacht 3 hours ago [-]
Truffles come to mind.
dice 19 hours ago [-]
I missed the word "Treat" in the title initially and I was incensed. All dogs are the best dog TYVM.
thih9 17 hours ago [-]
My dog told me to write that the set of treats is missing non bleached rawhide, other dehydrated meats (eg rabbit, goat, fish), animal parts (eg ears), and vegetables. Also, he volunteers.
buildsjets 14 hours ago [-]
My Husky used to CRAVE little dehydrated minnows. Extra crunchy.
abofh 10 hours ago [-]
This just feels like an excuse to give your dog a statically significant amount of treats
pike00 15 hours ago [-]
Interesting read! I might replicate with my two whippets. What’s their inter-rater reliability?
You also should continue with a swap of the hands; randomize which one is in the left and record the results to see if the left bias is real
jnellis 14 hours ago [-]
My dog's favorite treat is the most expensive one. Or the one that causes the most intestinal (and therefore cleanup) issues. All it takes is a pause and some brief math to see that a lot of (good for your dog) dog treats are $16/lb. That's not even remotely the most expensive. It's cheaper to feed them straight meat. I've resorted to making my own treats. It's like $4/lb, plus you know exactly what goes into it (mostly ground turkey, yams, rice/chickpea flour).
In my experience, it's usually a partly-rotten deer leg found in the bushes on a nature walk.
CM30 18 hours ago [-]
This makes me wonder what the research is for whether certain types or breeds of dog prefer certain dog treats, and how individual dogs might develop a preference for one kind over another. Based on this experiment it doesn't seem like the type of meat matters much, since while the top ranked treat is chicken, his second favourite seems to be the duck one.
7 hours ago [-]
doglover737484 17 hours ago [-]
Add some interesting smells and dogs will eat absolutely anything.
Used tampons, literal shit, soiled underwear, dierhia, dead bird, freshly killed cat, they owners... It is basically a pig with collar that lives in your house.
ngai_aku 15 hours ago [-]
* varies by breed
I've seen many a lab eat their own shit, but my standard poodle takes his time investigating even a piece of steak before he decides whether it's worth eating.
embedding-shape 10 hours ago [-]
It'd be hard to sleep with a pig in the bed with us though, but point taken. Now I kind of wanna get a pig-buddy for our dog.
dgacmu 15 hours ago [-]
And so useful when you don't feel like scraping the dishes. ;-)
dieselgate 16 hours ago [-]
Our Malinois really liked the dehydrated chicken too, currently finishing a bag of the Trader Joes organic chicken jerky sticks. Interesting idea for some experiments, thanks for sharing
Incidentally OP if you want to make it more adaptive, you can just fit the B-T model each time, and grab a posterior sample of what the best pair is, and test that, which turns out to be Thompson sampling. I did this for fun with blind taste-testing of mineral waters: https://gwern.net/water
Below, I filtered for A vs E, the top two choices. Notice how they switch left and right hand each time:
A/E :: A
E/A :: A
A/E :: E
E/A :: E
A/E :: E
The initial set-up was to do 3 comparisons of the 5 treats (30 trials), alternate between right and left hand, then write a quick python script to randomize the order.
A bit more than halfway through the experiment, I ran the model and realized that A/D/E were the only contenders left, so I removed the B/C trials and added more A/D/E trials.
Why would you give your dog those as treats?
You also should continue with a swap of the hands; randomize which one is in the left and record the results to see if the left bias is real
Used tampons, literal shit, soiled underwear, dierhia, dead bird, freshly killed cat, they owners... It is basically a pig with collar that lives in your house.
I've seen many a lab eat their own shit, but my standard poodle takes his time investigating even a piece of steak before he decides whether it's worth eating.
https://growlersdogbones.org/
Big Strawberry == Driscoll's, at least on the west coast. And those are some pretty undistinguished strawberries.
https://www.bigchicken.com/about