Chat GPT field scoring

[TradLife406]

After a lifetime of points I finally drew a big bull rifle tag for here in Montana.
I've killed a few bulls but in general districts it's usually just been whatever I could kill with the trad bow.

But with this tag there's a chance for a legit monster. And I have very little experience in guessing score.

Not sure if anyone has tried this with a phone skope. But it seems the Chat GPT app can score really close. I had it score my buck last year and it was pretty close.

So I googled for a 380" bull and uploaded it

Obviously you'd need service but just another tool in the tool box. Thought I'd share.

 

[ResearchinStuff]

i think it pulled you a verbatim answer from a field judging guide. the text is awfully familiar to something i read years ago.

 

[blkqi]

Love the enthusiasm

 

[mustelid_master]

With verbage like that, it should start making youtube videos of it's hunts with a single adjective in all caps as it's title

 

[LongWayAround]

I have Eastman's elk book with a chapter on field judging. I would be happy to share some pages. Send me a PM, if you like.

 

[cnelk]

Pro tip when using ChatGPT and want to to Copy and paste

Scroll all the way to the bottom of the message

Select this option and it will copy the message (instead of doing screen shots)

 

[voltage]

ChatGPT guessed 320-330 on my 2023 bull from one spotter pic. It went 323

 

[kodyshriver]

Hot damn this thing is good. Never saw the shed tine behind the buck either!

 

[rookieforever33]

Hot damn this thing is good. Never saw the shed tine behind the buck either!

Thats like the funhouse mirror that makes me taller and thinner. Except its wider, taller, and smiling for the camera

 

[kodyshriver]

Thats like the funhouse mirror that makes me taller and thinner. Except its wider, taller, and smiling for the camera

I asked it to enhance the smile

He was rutting a doe—I think this makes it even more accurate.

 

[atmat]

ChatGPT guessed 320-330 on my 2023 bull from one spotter pic. It went 323

Do you have the picture you used? Would like to see

 

[sz28]

After a lifetime of points I finally drew a big bull rifle tag for here in Montana.
I've killed a few bulls but in general districts it's usually just been whatever I could kill with the trad bow.

But with this tag there's a chance for a legit monster. And I have very little experience in guessing score.

Not sure if anyone has tried this with a phone skope. But it seems the Chat GPT app can score really close. I had it score my buck last year and it was pretty close.

So I googled for a 380" bull and uploaded it

Obviously you'd need service but just another tool in the tool box. Thought I'd share.

Are you sure it wasn't inspecting the metadata of the photo? I only ask because it mentions matching the "label". It also grossly overestimated the size of a bull I killed by 100" after I gave it two pictures to look at from two different angles.

 

[voltage]

Do you have the picture you used? Would like to see

 

[atmat]

View attachment 868842

So it accurately field judged based on this blurry photo? That’s pretty incredible.

 

[voltage]

So it accurately field judged based on this blurry photo? That’s pretty incredible.

Agreed.

 

[DarthVader]

chatGPT was able to score this california buck using a video of which i never was able to get a unobstructed full view of its antlers. pretty impressive. scored it at 140". pretty amazing.

 

[bz_711]

I tried it on some of my bulls...ChatGPT way lower than what I think they score


Joking aside I've never scored any of them...not sure I ever will.

 

[AZ_Hunter_2000]

Played around with this in May on known-score deer. It was bad and would leave folks disappointed. Hopefully it got better in the past two months.

Just now…
Gave it a lot of photos, field and wall, to score a 370 class bull that is on my wall. It gave a range of 320-340. If someone passed on the bull due to the score, they’d have missed out. If they shot it based on the score, they’d be pleasantly surprised.

Black Bear: one was 1” too big and the other was 1.75” too big.

Fun to play with but may not be the wisest option to make a shoot / don’t shoot decision.

From May…
Now what is interesting on ChatGPT, I asked it to score all of my scored Coues and mule deer. This is submitting multiple photos of each, different angles, and specifying the species. The most "accurate" score was 10" too high, had one mule deer scored over 40" too high (being told its a 200" but is actually 160"), and a Coues that was scored over 50" too high (a 138" Coues that really scores 88" - way cool 5x5 though). The rest all scored too high by 15-30". It was interesting that not a single buck scored by ChatGPT came in lower than the actual score; 100% were above the actual score. I have not asked it score any of my other game species.

 

[acmckeage03]

Just used it to score one of my bulls and it’s off 20-40” this bull falls in the 300” range it’s telling me 320-340.

That’s a fantastic bull elk—congrats! Based on the photos, here’s a rough field estimate of the score using the Boone & Crockett (B&C) scoring method for typical American elk:








Main Estimations (based on visual inspection):

• Main beam length (each side): ~50–52 inches
• Tine lengths:
  • G1 (eyeguard): ~16–18”
  • G2: ~14–16”
  • G3: ~12–14”
  • G4 (royal): ~12–14”
  • G5: ~8–10”
  • Possibly a G6 or crown: ~4–6”
• 
  
• Inside spread: ~38–40”
• Circumference measurements (H1–H4): Averaging ~6–7” each

Estimated Total:

• Main beams: ~104” (52” x 2)
• Tines (12 total): ~130”
• Spread: ~38”
• Mass (8 measurements): ~50–55”

Estimated Gross Score: ~320–340 B&C

• If the antlers are symmetrical with minimal deductions, the net typical score might fall in the 310–330 range.

If you want a more accurate score, you can tape it using the B&C scoring sheet—I’d be happy to walk you through that process if you want.





Let me know if you’d like me to generate a diagram or overlay to visualize the measurements on the antlers.

 

[duckcarver]

This is interesting. I just let ChatGPT run with photos of an elk and a sheep. Both estimates were within 5-9% of the taped scores.