Bondage Element Combination Limit Test: How Reproduction Rate Changes from 2 to 3 to 4 Elements

Key Findings

2 elements reproduce reliably. Both blindfold + hands tied (Condition A) and blindfold + ball gag (Condition B) showed high reproduction rates
3 elements (Condition C) still reproduced all elements in 3 out of 3 images. However, the “tightness” of the hand binding started to vary
4 elements (Condition D) caused outfit collapse. The specified black lingerie was replaced by leather harness gear, with bondage elements interfering with the clothing prompt
In practice, up to 3 elements can be reliably reproduced. At 4 or more, the risk of unexpected outputs due to element interference is high

Experiment Design

This test progressively increases the number of bondage elements specified in a single prompt to find the practical reproduction limit.

Conditions

Condition	Elements	Bondage Elements
A	2	Blindfold + Hands tied
B	2	Blindfold + Ball gag
C	3	Blindfold + Ball gag + Hands tied
D	4	Blindfold + Ball gag + Hands tied + Collar & Leash

Common Settings

Subject: 1girl, 32yo japanese actress
Outfit: black lingerie
Pose: kneeling on bed
Lighting: soft lighting
3 images per condition (fixed seeds)

Condition A: Blindfold + Hands Tied (2 Elements)

Condition A: blindfold + hands tied

1girl, 32yo japanese actress, black blindfold, hands tied behind back with silk ribbon, black lingerie, kneeling on bed, soft lighting

NSFW - クリックで表示

seed6000	seed6001	seed6002

Blindfold: Reproduced in 3 out of 3 images as a black cloth blindfold.

Hands tied: Silk ribbon was confirmed near the wrists in all 3 images. However, in seed6001 and seed6002, the ribbon appeared loosely wrapped rather than tightly binding the wrists. Seed6000 gave the strongest impression of actual restraint.

Outfit interference: None. All 3 maintained black lingerie.

Condition B: Blindfold + Ball Gag (2 Elements)

Condition B: blindfold + ball gag

1girl, 32yo japanese actress, black blindfold, red ball gag in mouth, black lingerie, kneeling on bed, soft lighting

NSFW - クリックで表示

seed6010	seed6011	seed6012

Blindfold: Reproduced in 3 out of 3.

Ball gag: Red ball gag was reproduced in the mouth in all 3 images. Both color and shape were highly consistent.

Outfit interference: None. Black lingerie was maintained.

Both 2-element combinations (A and B) were stable with no interference between elements.

Condition C: Blindfold + Ball Gag + Hands Tied (3 Elements)

Condition C: blindfold + ball gag + hands tied

1girl, 32yo japanese actress, black blindfold, red ball gag in mouth, hands tied behind back with silk ribbon, black lingerie, kneeling on bed, soft lighting

NSFW - クリックで表示

seed6020	seed6021	seed6022

Blindfold: Reproduced in 3 out of 3.

Ball gag: Red ball gag reproduced in 3 out of 3.

Hands tied: Ribbon was confirmed near the wrists in all 3 images. In seed6020, the ribbon wrapped around the body in a binding pattern. Seed6022 showed the clearest behind-the-back binding. Seed6021 had ribbon present at the hands but with weaker restraint impression.

Outfit interference: None. All 3 maintained black lingerie.

All 3 elements were simultaneously reproduced, though the “clarity of restraint” for the hand binding varied across seeds.

Condition D: Blindfold + Ball Gag + Hands Tied + Collar & Leash (4 Elements)

Condition D: blindfold + ball gag + hands tied + collar

1girl, 32yo japanese actress, black blindfold, red ball gag in mouth, hands tied behind back with silk ribbon, leather collar with leash, black lingerie, kneeling on bed, soft lighting

NSFW - クリックで表示

seed6030	seed6031	seed6032

Blindfold: Reproduced in 3 out of 3.

Ball gag: Reproduced in 3 out of 3.

Hands tied: Only seed6031 showed clear behind-the-back binding. In seed6030, the restraint morphed into a harness-style bondage gear with unclear wrist binding. Seed6032 showed wrist restraints but they were no longer “silk ribbon.” Clear reproduction in 1 out of 3.

Collar: Black leather collar confirmed in all 3 images (3 out of 3).

Leash: All 3 images showed strap-like elements, but none clearly depicted an independent “leash.” The leash concept merged with harness straps.

Outfit interference: Severe interference occurred. In seed6030, the black lingerie was completely replaced by a leather harness, producing an entirely different outfit than specified. Seed6032 generated unprompted stockings. The increased bondage elements pulled the entire style toward “bondage gear.”

Element-by-Element Reproduction Summary

Element	A (2 elem.)	B (2 elem.)	C (3 elem.)	D (4 elem.)
Blindfold	3/3	3/3	3/3	3/3
Ball gag	-	3/3	3/3	3/3
Hands tied	3/3	-	3/3	1/3
Collar	-	-	-	3/3
Leash	-	-	-	0/3
Outfit preserved	3/3	3/3	3/3	0/3

Blindfold and ball gag remained stable regardless of element count (likely because they are visually prominent items on the face)
Hands tied reproduced up to 3 elements but tended to be absorbed into other restraint gear at 4 elements
Leash failed to reproduce as an independent element, merging with harness straps
At 4 elements, the outfit collapsed and the entire output was pulled toward a bondage gear aesthetic

Note: Two elements reproduce without issue, and three elements still work, but four elements tend to produce chaotic output. Blindfold and ball gag stay stable regardless of element count, while hands-tied is the first to be dropped — face-area items appear to be prioritized. At 4 elements, seed6030 produced a full leather harness despite specifying lingerie, with the output pulled toward the “bondage” concept. In practice, 3 elements appears to be the upper limit; splitting the additional elements across two generation passes may be a more reliable approach.