Why 45 Video Prompts and About 200 Final Shots Are Both Correct for the Same Film

Borrowed Wounds has 45 nonempty video prompts and about 200 final shots.

That is not a contradiction. It is a warning about units.

The production board contains 50 records for the film. Of those, 45 contain a nonempty shot prompt. The finished edit produced 208 detected segments, but calibration put the defensible estimate at about 198 final shots. That is why I describe the film as having around 200 shots.

If I place 45 and about 200 in a column labeled “shots,” one of them has to look wrong. The problem is the column.

The prompt count belongs to planning

The 45 figure comes from the production board. It counts nonempty shot prompts.

That tells me how much written shot direction exists in that structured source. It does not tell me how many final cuts the film contains. It also does not establish the total number of generated clips, discarded attempts, or retries.

Even the 50 production records cannot be relabeled as 50 video prompts. Only 45 contain nonempty shot prompts.

The same problem appears across the wider set of projects. The boards contain 515 production records and 341 nonempty shot prompts. Those 515 records include characters, supporting characters, places, props, and wardrobe. They are not all video directions, and they are not all character records.

I also wrote 1,078 image prompts. Added to the 341 shot prompts, that produces 1,419 structured prompts. Only 713 of the image prompts have a current, verifiable image output.

That larger inventory makes the category problem easier to see. Written records, prompts with content, current image outputs, and final shots belong to different stages. Adding them may produce a correct arithmetic total, but that total does not become a useful production unit by itself.

The final-shot count belongs to the edit

The around-200 figure comes from the finished film, not the prompt board.

Shot detection produced 208 segments. Calibration estimated about 198. I therefore describe the film as having “around 200 final shots” rather than treating 208 as exact.

That difference matters. A detector returns a result. Calibration decides how much confidence I should place in that result. Reporting 208 as the exact shot count would ignore the adjustment I had already made.

Reporting 198 as perfectly exact would also push the evidence too far. It is an estimate.

“Around 200” is less sharp on the page, but it is the more honest count.

This is the number I use when discussing the structure of the finished edit. It is not the number I use when discussing how many shot prompts were written.

The missing number sits between them

I did not log a total count of all generated video shots, including discarded attempts and retries.

That missing total prevents me from turning the gap between 45 prompts and about 200 final shots into a generation estimate. I cannot use the final edit to work backward and claim a clip count I never logged.

I also cannot assume that every written prompt produced one final shot. The two verified counts already show that a prompt count and an edit count do not share the same unit. The counts establish the difference. I do not have a complete conversion rule between them.

This is where a budget can become fiction.

If I price a project by prompt count, I am measuring written instructions.

If I price it by generated clips, I need a clip-generation count that I did not log.

If I measure the delivered edit by final shots, I use the calibrated estimate of around 200.

Collapsing those categories into one “video count” hides the missing evidence. It can make planning look cheaper or make delivery look impossible, depending on which count someone treats as the baseline.

Delivery is its own threshold

I cannot verify a mounted film for any of the other festival projects.

They still have written shot prompts. Copycat has 14. Everybody Sleeps at Nine has 20. Low Water has 21. Sacar las maletas has 24. Set Him Down has 21. The Lord Is With Thee has 150. Three Seconds has 8. Unclaimed has 8. Unravelled has 15. What Darkness Keeps has 15.

Those counts prove structured preparation. They do not prove final shots because I cannot verify a mounted film for those projects.

This is why I resist using prompt volume as a proxy for delivered footage. The Lord Is With Thee has 150 written shot prompts, more than the 45 recorded for Borrowed Wounds, yet I cannot verify a mounted film to count for it.

The prompt is real work. The final shot is real work. They are simply evidence of different things.

For Borrowed Wounds, I can defend 45 as the number of nonempty video prompts in the board. I can defend around 200 as the calibrated estimate of shots in the final edit. I cannot defend a total for generated clips, discarded attempts, and retries because I never logged that total.

Today, add separate columns to your own production ledger for written prompts, generated clips, selected edit segments, and final shots. If one count is missing, mark it unknown. Do not make another column impersonate it.

---

Credits spent, attempts counted, results measured. I trust numbers because they don't flatter anybody, least of all me.