Dolphin Encounter Rates

Looking through 310 unique Minecraft 1.16.1 ranked matches, dolphins appeared before first Nether entry in roughly one out of every five usable player replays.

The overall measured dolphin encounter rate was:

112 / 559 = 20.0%

Shipwreck and buried treasure routes were close:

Shipwreck:       20.7%
Buried treasure: 19.4%

The larger difference was between players in the same match.

Out of 253 matches with usable data from both runners, both players encountered dolphins only 8.7% of the time. Both players encountered their first dolphin in the same chunk only 2.4% of the time.

What counted as a dolphin encounter

Each player replay was treated as one trial.

The measurement window was:

first valid Overworld position
        ↓
shipwreck or buried treasure
        ↓
first Nether entry (magma ravine portal entry)

A replay counted as positive if at least one dolphin was recorded within 24 three-dimensional blocks of the player during that window.

It does not necessarily mean:

• the dolphin was seen
• the dolphin was used
• that dolphin’s grace was received

Why Dolphin’s Grace wasn’t tracked

The replay data records player movement, swimming pose, and nearby dolphin positions, but it does not record status effects directly. Fast swimming near a dolphin could suggest Dolphin’s Grace, and it is probably detectable from swim state + velocity, but I didn’t end up checking for it in this analysis. Because of that, the data only tracks whether a dolphin came within 24 blocks. It does not try to classify whether the runner actually received or used Dolphin’s Grace (also sometimes people just don’t see the dolphin).

Dataset size

I used a bunch of recent matches from top players to get replays for this dataset, and some of the matches overlapped between players, so duplicate match/player combinations were removed before calculating rates (they should all be unique matches).

After deduplication:

The basic trial split was:

559 valid player runs
├── 112 dolphin
└── 447 no dolphin

The 61 incomplete timelines were excluded rather than counted as no-dolphin runs.

Why timelines were excluded

All 61 excluded player timelines loaded successfully and contained normal player and entity data. Each one had at least one Ranked reset marker:

61 excluded player timelines
├── 54 with one reset
├──  6 with two resets
└──  1 with three resets

A reset splits the replay into separate attempts. I chose to reject these because combining attempts could create a false dolphin result, while treating only part of the recording as a complete no-dolphin run could bias the rate. The opponent’s uninterrupted timeline was still usable in 53 of the 57 affected matches; both players reset in the other four.

Dolphin encounter rate

Shipwreck

Buried treasure

Combined

Conceptually:

Shipwreck
57 dolphin / 275 trials
        ↓
20.7%

Buried treasure
55 dolphin / 284 trials
        ↓
19.4%

The measured shipwreck rate is 1.4% higher, but the confidence intervals overlap heavily. This dataset does not provide a strong reason to treat the two route types as having different underlying dolphin rates.

The useful estimate is the pre-Nether dolphin encounter rate = ~20%.

Detection distance (classification)

This isn’t very useful since 24 blocks is the replay-entity tracking limit, but I thought I would include it anyways.

For the first qualifying dolphin in each positive trial:

This happens because the replay records the dolphin as it enters the nearby entity-tracking radius. A dolphin already moving around outside the radius generally appears in the data when it crosses that boundary.

The actual dolphin may have:

• spawned farther away
• already existed before the runner approached
• moved toward the runner
• been brought into range by the runner’s movement

The useful statement is only that the dolphin came within 24 blocks before the runner entered the Nether.

Number of dolphins per positive run

Some positive trials contained more than one tracked dolphin.

Most positive runs therefore had one or two dolphins:

112 dolphin-positive runs
├── 59 with one dolphin
├── 39 with two dolphins
├──  8 with three dolphins
├──  5 with four dolphins
└──  1 with five dolphins

Minecraft 1.16.1 dolphin spawn entries use groups of 1–2, so larger observed counts can represent multiple groups entering range rather than one unusually large natural pack.

What happened between both players

The more relevant result for ranked parity comes from matches where both player perspectives were usable.

There were 253 such matches.

In tree form:

253 complete player pairs
├── 177 neither player
├──  54 exactly one player
├──  16 both, different chunks
└──   6 both, same chunk

At least one player encountered a dolphin in 76 matches:

76 matches with at least one dolphin
├── 54 only one player
└── 22 both players

That means 71.1% of matches with any dolphin encounter gave it to only one runner.

Natural dolphin spawning was therefore much more likely to produce a mismatched result than a shared one.

Same-chunk encounters

Only six of the 253 complete matches gave both players their first dolphin in the same chunk.

The tick differences for those same-chunk detections were:

tick difference: 1
tick difference: 14
tick difference: 20
tick difference: 65
tick difference: 357
tick difference: 959

The first three happened close together, while the later three show that the same first-dolphin chunk does not necessarily mean both runners encountered the same dolphin group.

Different-chunk encounters

The 16 both-positive matches with different chunks looked different:

The large difference in ticks and chunks in these encounters make it likely that the spawns were quite independent. Both players having a dolphin somewhere in the route does not usually mean they received the same spawn (since standardization isn’t implemented for dolphins as of v5.8.23).

Shipwreck vs. buried treasure pairs

The paired outcomes can also be separated by route.

Shipwreck

Buried treasure

The paired results are also very similar between seed types, and the current data doesn’t show a reliable difference between the two.

What “natural rate” means

This is observed MCSR Ranked route data rather than a random sampling of vanilla ocean biomes. Vanilla biome tables only describe the conditional weight of choosing a dolphin when the game successfully reaches the relevant water-creature spawn selection. They do not directly give the chance that a runner encounters a usable dolphin during a route since they are only exposed to a limited number of chunks, specific mob counts (squid and other creature competition), terrain differences, valid water positions, travel time, spawn-cycle timings, etc.

So by using actual run data from MCSR matches, we can get a more realistic estimate.

Sample limitations

The combined ~20% estimate still has a 95% confidence interval of 16.9%–23.6%.

Different runners take different lines at different times, at various render distances, and that’s part of the real encounter process and can skew the results.

Summary (TLDR)

From 310 unique MCSR Ranked 1.16.1 matches:

559 valid player trials
├── 112 dolphin
└── 447 no dolphin

Measured encounter rates:

Shipwreck:       57 / 275 = 20.7%
Buried treasure: 55 / 284 = 19.4%
Combined:        112 / 559 = 20.0%

Detection distance:

threshold: 24 blocks
mean first detection:   23.66 blocks
median first detection: 23.72 blocks
observed range:         22.77–24.00 blocks

(see detection distance section for more information)

Paired results:

253 complete matches
├── neither player:              70.0%
├── exactly one player:          21.3%
├── both, different chunks:       6.3%
└── both, same first chunk:        2.4%

The shipwreck and buried treasure rates are not meaningfully different from each other with the current sample size.

Sources

• MCSR Ranked Source environment for the Minecraft 1.16.1 ranked match and replay data.
• Minecraft Wiki — Dolphin General reference for dolphin spawning and behavior.