Interested in joining the Silph Research group? Learn More »

Silph Study: #059


Demystifying Ditto - Rarity and Shiny Rate

Ditto — the Transform Pokémon — has been one of the most elusive and least understood encounters in Pokémon GO since its release all the way back in November 2016. Because Ditto typically appear in disguise as another Pokémon and the available disguises change frequently, travelers are frequently unaware of when and how often they encounter a Ditto in the wild. On top of that, its rarity and low catch rate make Ditto one slippery mark to capture.

In today’s article, we Defog some of the mysteries behind Ditto. A small group of Silph researchers spent five months tracking down over 13,000 Ditto disguises and have made some transformational discoveries. Let’s see what they have unveiled!

Key Findings

  • Around 2% of all Pokémon disguise species encountered in the wild are Ditto.
  • Ditto likely uses the permaboosted shiny rate of 1 in 64.

Background and Methodology

Ditto takes the form of one of a small group of Pokémon disguises, appearing as the disguise both on the map and in the encounter screen. To work out how many of those possible disguises were secretly Ditto, Silph Researchers attempted to catch every eligible disguise they encountered until it was caught or fled. They recorded how many disguises they caught and how many they saw as shown in their Pokédex entries. Researchers also detailed how many of those disguises turned into Ditto, the number of shiny disguises, and the number of shiny Ditto that they found on their travels.

From these Pokédex numbers, we were able to determine the total number of encounters that fled. We applied a flee correction method that proportionally splits the number of flees between Ditto and disguises based on the Base Catch Rate (BCR) and Base Flee Rate (BFR) for each species. This gave us an indication of the total number of disguises — and number of Ditto — our researchers came across while out in the field. For more details on our data collection and flee correction methods, see the appendix.

Ditto Frequency

Our primary goal in this study was to determine the “Ditto rate”. We define Ditto rate as the proportion of eligible disguises that are secretly Ditto. Our study found the overall Ditto rate across all available disguises to be:


with a 95% confidence interval (CI) of 1.78 to 2.25%.

At 1 in 50 odds, this makes Ditto quite a rare encounter indeed! Ever the avid Ditto hunters, we were curious if there might be a trick or strategy to improve the odds of a Ditto encounter. We examined several factors, and here’s what we discovered.

Ditto Rate Across Species

After determining the overall Ditto rate, we examined the rate for each disguise species that had met its goal of 600 or more observations (Drowzee, Ekans, Gastly, Lillipup, Natu, and Surskit). The graph below shows the rate we observed for each disguise.

We failed to reject the null hypothesis that Ditto was found in equal proportions across the tested disguises (χ²(5) = 4.2, p = 0.52). In general, this means each disguise has the same chance of turning into Ditto. Might this change during events featuring either Ditto or one of its disguises?

Disguise and Ditto Featured Events

Over the course of our study, we examined various in-game events and how they might affect the Ditto rate. To start, we looked at events that featured one or more disguises but did not specifically feature Ditto. Though disguises were much more common during these events, the rate at which they transformed into Ditto remained unchanged at roughly 1 in 50.

This was not the case for events that explicitly featured Ditto, such as The April Fools’ 2-Oh?-22 event. The official press release announced that Ditto would be transforming into different disguises in the wild. In addition, the “A Ditto Surprise” portion of the event (5 pm – 8 pm local time) featured Ditto disguises appearing more frequently in the wild. Our study found that the Ditto rate during “A Ditto Surprise” was a staggering 49% (95% CI [46.54, 51.73%])! The rest of the April Fools’ event (excluding the 3-hour spotlight) had a rate of 9% (95% CI [5.21, 14.21%]).

Finally, we have the June TCG event, which featured three additional Ditto disguises: Numel, Spinarak, and Bidoof. During the event, Researchers collected data for these added disguises as well as the existing Ditto disguises: Natu, Lillipup, Swirlix, and Gastly. The table below shows the Ditto rate for each disguise species.

Disguise Species Number of Disguises Number of Ditto Ditto Rate
[w/ 95% CI]
Numel 772 213 21.63% [19.14, 24.28%]
Spinarak 663 169 20.26% [17.63, 23.09%]
Bidoof 543 131 19.47% [16.61, 22.59%]
Spinarak (Spotlight Hour) 928 221 19.22% [17.02, 21.57%]
Natu 339 5 1.55% [0.61, 3.29%]
Lillipup 262 10 3.50% [1.78, 6.20%]
Swirlix 9 0 0.00% [N/A]
Gastly 8 0 0.00% [N/A]

Our study indicated that the featured disguises had a combined Ditto rate of 20.16% (95% CI [18.88, 21.48%]). Interestingly, we found that only the additional featured disguises of Numel, Spinarak, and Bidoof had the higher Ditto rate. The seasonally available disguises like Natu and Lillipup kept their standard Ditto rate. In fact, this was the only instance throughout our study where different disguises had different Ditto rates!

Shiny Ditto

Shiny Ditto was introduced with the February 2021 Pokémon GO Tour: Kanto event, where upon completion of the first Special Research story, all paid ticket holders received a shiny blue research Ditto. After the GO Tour, Ditto’s shiny was disabled until September 2021, when it featured in the month’s research breakthrough and its shiny was re-enabled. Simultaneously, shiny Ditto was made available in the wild for the first time with a refreshed list of disguise species.

To understand Ditto’s shiny rate, Silph researchers tracked the number of normal and shiny Ditto caught across all disguise species. After careful data collection, we believe that the probability of receiving a shiny Ditto is approximately:

1 in 64

(the exact observed rate was 1 in 56.5 with a 95% confidence interval of 1 in 81.9 to 1 in 40.3).

This means Ditto falls within the permaboost category of our shiny database.

Ditto Shiny Rate Across Events

In addition to examining the overall Ditto shiny rate, we looked at Ditto featured events to see if the shiny rate might have changed. There was no difference in the shiny rate between Ditto caught outside of Ditto featured events and those caught during Ditto featured events.

Period Type Ditto Caught Shiny Ditto Rate (1 in X) 95% CI
Standard Ditto Rate 247 3 1 in 82.3 [1 in 291.3 to 1 in 31.2]
April Fools’ 745 15 1 in 49.7 [1 in 84.6  to 1 in 31.1]
TCG Event 704 12 1 in 58.7 [1 in 106.9 to 1 in 34.9]
Total 1625 30 1 in 56.5 [1 in 81.9 to 1 in 40.3]

As the permaboost rate falls within the 95% confidence interval for each period, we have no reason to believe that Ditto’s shiny rate was altered during events, even if the Ditto-per-disguise rate was boosted.

Parting Words

We hope this study helps you in your hunt for Ditto. Ditto is quite the rare encounter, and even with a permaboosted shiny rate, on average we expect you would need to catch 3,200 disguises before one transformed into a shiny Ditto!

One advantage we’ve shown is that Ditto appears to have the same chance across its disguises. This likely indicates that Ditto status is determined after the disguise spawns. Searching for disguises that are common in certain biomes, spotlight hours, or events will give you more chances to roll a Ditto.

Specific Ditto featured events, such as April Fools’, provide some of the best opportunities to encounter Ditto, as they can increase the overall Ditto rate by 10-fold or more. In fact, during our study, the vast majority of Ditto and their shiny variants were caught during such events.

That’s it for now, travelers, see you — or your disguise — on the road!

Authors: Scientist CaroKann, Lead Researcher Belle, Senior Researchers Cloudstarcarrie and ThatOneZebra
Analysis: Lead Researchers Tobias and Belle, Senior Researchers Cloudstarcarrie and ThatOneZebra, and Scientist CaroKann
Project Leaders: Lead Researchers Brush are Best, PancakeIdentity, and turtleduck
Graphics: Scientists WoodWoseWulf and Senior Researcher Cloudstarcarrie
Editing: Scientists Cham1nade and Titleist, and Lead Researcher JinianD

The data for this article was analyzed by a unique team of rookie and veteran analysts working together to transmit the tricks of the trade. The following contributed their time and effort to make this article possible:

BGreen017, bowlergirlj, careatheadrest,
Cloudstarcarrie, grundgesetze, jetigig,
Lauracb18, marinaj, natanos,
RachelBatman, SpelingChanp, ThatOneZebra,
Tober, Tobias, topseeser,
TrisChandler, turtleduck, and Willie333b

Appendix 1: Ditto Mechanics Background

During an encounter with a Ditto, the disguise will visually show its own weather boost tag. For example, a Natu disguise will display the windy weather tag in windy weather. The deception is further maintained through the catch circle, which will mimic that of the disguise and does not reflect Ditto’s lower Base Catch Rate. Even though a Ditto encountered in the wild will be a Ditto for every trainer who catches it, we were unable to use this feature to recheck fled Disguises within the constraints of this study.

One key difference between an encounter with a Ditto and its disguise is that Ditto is often harder to catch than the Pokémon it is disguised as. With a low Base Catch Rate (BCR) of 20%, Ditto is much more likely to break out of the ball than most of its disguises. Every time Ditto breaks out, it has a chance of running away. Ditto’s Base Flee Rate (BFR) – its chance to run away – is 10%. Our flee correction method used this knowledge of BCR and BFR to predict how many flees were Ditto.

Although Ditto doesn’t nest, its disguises can. Some disguises may only be available in the Northern or Southern hemisphere, or during specific events. Usually, Ditto can disguise itself as five to nine Pokémon species at a given time. The table below lists when each disguise was available or featured in an event over the course of our study.

Disguise Wild and Event Availability Timeline
Stunky 2022/03/01 – 2022/03/31 (Seasonal Spawn – Southern Hemisphere)
Remoraid 2022/03/01 – 2022/03/31 (Seasonal Spawn)
Numel 2022/03/01 – 2022/03/31 (Seasonal Spawn)
2022/06/16 – 2022/06/30 (TCG Crossover Event)
Gastly 2022/03/01 – 2022/08/31 (Seasonal Spawn)
2022/04/01 (April Fools’)
2022/04/28 – 2022/05/02 (A Mega Moment)
2022/06/16 – 2022/06/30 (TCG Crossover Event)
Foongus 2022/03/01 – 2022/03/31 (Seasonal Spawn – Northern Hemisphere)
Drowzee 2022/03/01 – 2022/03/31 (Seasonal Spawn)
2022/03/14 – 2022/03/21 (Festival of Colors)
Swirlix 2022/04/01 (April Fools’)
2022/06/01 – 2022/08/31 (Seasonal Spawn)
2022/06/16 – 2022/06/30 (TCG Crossover Event)
Surskit 2022/04/01 (April Fools’)
2022/04/01 – 2022/08/31 (Seasonal Spawn)
2022/05/12 – 2022/05/20 (Water Festival 2022)
Natu 2022/04/01 (April Fools’)
2022/04/01 – 2022/08/31 (Seasonal Spawn)
2022/06/16 – 2022/06/30 (TCG Crossover Event)
Lillipup 2022/04/01 (April Fools’)
2022/04/01 – 2022/08/31 (Seasonal Spawn)
2022/06/16 – 2022/06/30 (TCG Crossover Event)
Finneon 2022/04/01 (April Fools’)
2022/04/01 – 2022/06/01 (Seasonal Spawn)
Ekans 2022/04/01 (April Fools’)
2022/04/01 – 2022/06/01 (Seasonal Spawn)
Dwebble 2022/04/01 (April Fools’)
2022/06/01 – 2022/08/31 (Seasonal Spawn)
2022/06/12 – 2022/06/12 (Shieldon and Cranidos Research Day)
2022/06/16 – 2022/06/30 (TCG Crossover Event)
Bidoof 2022/06/16 – 2022/06/30 (TCG Crossover Event)
Spinarak 2022/06/16 – 2022/06/30 (TCG Crossover Event)
2022/06/21 (TCG Spotlight Hour)

Appendix 2: Detailed Data Collection Methods

Data for this study was collected between March 10 and August 10, 2022. Before recording any encounters, researchers declared in advance which disguise species they would like to track and for how long. Observation periods lasted up to seven days. At the end of each observation period, researchers had to submit the numbers of disguise species and Ditto observed before they could declare a new observation period. Researchers also declared if the species they were tracking was featured in an event or a spotlight hour. Declarations were closed before and after any events featuring Ditto or a disguise to ensure that these observation periods were not mixed with non-event periods.

In order to count how many disguises and Ditto a researcher encountered, we relied on Pokédex entry screenshots taken at the start and end of each observation period. When a disguise or a Ditto is caught, the ‘Number Seen’ and ‘Number Caught’ for the caught Pokémon’s Pokédex entry each go up by one. If the disguise flees, only the disguise’s Pokédex entry changes — the ‘Number Seen’ goes up by one, while the ‘Number Caught’ stays the same. In other words, when the disguise flees, Ditto’s Pokédex entry is unchanged.

However, if a Ditto flees while encountered using an autocatcher (such as a ‘Pokémon GO Plus’ or ‘Go-tcha’), its Pokédex entry does change — the seen count goes up by one, and the caught count is unchanged. In this scenario, the disguise’s Pokédex entry is unchanged, meaning we would have no way of knowing which disguise the Ditto had taken. For this reason, researchers did not use an autocatcher while collecting data for this study.

If a researcher clicks on a disguise and does not try to catch it (for example by escaping the catch screen or by using ‘quick-catch’ methods that do not result in a catch), the Pokédex entry for the disguise will change. The ‘Number Seen’ will go up by one, but the ‘Number Caught’ will remain the same. As this mechanic is indistinguishable from what happens when a disguise flees, researchers attempted to catch every disguise that they clicked on until it was finally caught or fled.

To determine the shiny rate for each shiny-eligible disguise species, we compared the number of shiny disguises encountered to the total number of disguises (with the flee correction). The observed shiny rate of shiny-eligible disguises was also compared with data from our regular shiny tracking study to ensure that the shiny rate had not changed since the species became a Ditto disguise. Once we were confident that the rate was the same, the data collected as part of this study was added to The Silph Road shiny rates page. Over the course of our study, we did not observe any instances where a Pokémon species’ shiny rate changed once it became an eligible Ditto disguise.

Although Niantic reassured trainers that shiny Pokémon would not transform into Ditto, travelers were quick to report instances where their shiny disguise did transform into Ditto:

  1. September 26, 2021, just over three weeks after the new mechanic was introduced, shiny Drowzee was encountered as a daily spawn and then transformed into Ditto.
  2. April 1, 2022 (April Fools’ Oh? event), a shiny field research disguise transformed into Ditto.

It’s likely that a variation in the encounter type (field research, daily spawn) led to these bugs. There have been no recent reports of a shiny disguise encountered in the wild transforming into Ditto. Even in cases where this mechanic behaves as intended, however, we do not know if the shiny disguise takes priority over the Ditto or if the Ditto takes priority over the shiny disguise when the Pokémon is spawned.

Given this uncertainty, we choose to ignore this interaction in our shiny rate and Ditto rate analysis. As most disguises use a base shiny rate, we would expect this interaction to happen about once in 25,000 encounters, and less than once in our 13,000+ observations.

Appendix 3: Detailed Flee Correction Method

Calculating the Ditto rate directly from observed catches only (ignoring flees) may provide some approximation of the true rate. However, this approach fails to account for a number of factors. The proportion of Ditto in a sample of caught Pokémon can differ from that in a sample of fled Pokémon, as Ditto is more likely to flee than its disguises, due to its low BCR. This results in Ditto being overrepresented in the number of fled Pokémon and underrepresented in the number of caught Pokémon. Determining the Ditto rate is further complicated by the variation in catching strategies among researchers. Because of these factors, we were unable to directly use the Ditto rate of caught encounters to determine the Ditto rate among fled encounters. Therefore, we needed to devise a method to estimate the proportion of fled Ditto vs disguises that took these factors into account.

Flee Correction Formula

If we could define a Total Catch Rate (TCR) based on the BCR and BFR of each Pokémon at different possible catch multipliers, we could determine the proportion of fled Ditto vs disguises. Our first step was to derive such an equation. Starting with the following:


Under the assumption that all researchers will attempt to catch a Pokémon until it is caught or flees, we can derive:

Therefore, the number of Pokémon (both Ditto and disguises) that flee breaks down to:

With this equation, all variables aside from the TCR are known, and the correlation between the two TCRs allows us to calculate an approximation of the true TCR_Ditto.

Flee Correction Table

To determine Ditto’s or a disguise’s TCR, we used their respective BCR and BFR as seen below.

Species Type Ditto
Gastly Natu Finneon
Ekans Lillipup
BCR 0.2 0.3 0.4 0.4 0.5 0.5 0.5
BFR 0.1 0.1 0.1 0.15 0.1 0.15 0.2

As the TCR will vary depending on catch multiplier (berry usage, throw accuracy, curveball, etc.), a TCR was calculated at each catch multiplier in the range {1.0, 1.1, 1.2, …, 8.0}. We simulated 106 Pokémon encounters at each BCR, BFR, and Multiplier combination using a random, uniform distribution of CPM values from Pokémon levels 1-30. (Our project did not track weather, but including CPMs from weatherboosted levels 31-35 had a negligible effect on the result.) The total number of catches and flees derived from these simulations were used to determine the TCRs for Ditto and its disguises.

Putting Everything Together

For each species, we modeled the total number of flees we would expect at each multiplier/TCR combination. We then selected the multiplier/TCR combination that best fit the observed number of flees. The number of Ditto caught and Ditto’s TCR at this multiplier was used to determine the number of flees which were Ditto. The remaining flees were assigned as disguise flees.

Testing our method

The numbers presented in this article rely on the accuracy of our flee correction method. To test how effective our method was, we ran various simulations on Ditto/disguise encounters at various Ditto rates. We found our flee correction method closely predicted the number of Ditto which fled in our simulations (usually well within a 15% margin of error for identifying the number of fled Ditto, and within 5% for identifying the total number of Ditto once we factored in known caught Ditto.) This held true even when our simulated catch attempts employed different catching strategies and berry usage on each throw.

While we have no way of going back and knowing exactly how many flees were actually Ditto, we believe that the numbers presented in the article are reasonably accurate. For context, outside Ditto featured events, over 98% of all encounters were caught by the researcher before fleeing. Supposing that every flee was actually a Ditto, our study would have a maximum Ditto rate of 3.85%. If none of the flees were Ditto, then we’d have a minimum 1.86% Ditto rate. At the very least, we believe the 2.00% Ditto rate presented in this article to be more accurate than those extremes. As we get more data and our models evolve, we may choose to revisit and/or improve the flee correction method presented here in future Ditto-related studies.

The table below shows the rounded numbers of disguises and Ditto caught before and after flee correction (excluding Ditto featured events).

Disguise Disguises Caught Ditto Caught Flees Disguises
Ditto Rate
[w/ 95% CI]
Lillipup 3876 78 86 3956 84 2.07% [1.67, 2.55%]
Surskit 2788 46 59 2844 49 1.71% [1.28, 2.23%]
Drowzee 1988 46 26 2010 50 2.41% [1.81, 3.14%]
Ekans 1480 31 28 1506 33 2.18% [1.53, 3.00%]
Natu 1305 22 46 1349 24 1.74% [1.14, 2.53%]
Gastly 688 12 14 701 13 1.80% [1.01, 2.98%]
Remoraid 168 4 2 170 4 2.47% [0.87, 5.61%]
Dwebble 146 1 0 146 1 0.68% [0.07, 3.13%]
Finneon 110 3 3 113 3 2.92% [0.90, 7.23%]
Swirlix 84 1 0 84 1 1.18% [0.13, 5.37%]
Numel 80 2 0 80 2 2.44% [0.51, 7.59%]
Stunky 25 0 0 25 0 0.00% [0.00, 9.47%]
Foongus 10 1 0 10 1 9.09% [1.00, 35.30%]

Across all event featured disguises, we found the overall Ditto rate to be 2.01% (95% CI [1.64, 2.43%]). This closely aligns with the non-event Ditto rate of 2.00% (95% CI [1.72, 2.31%]). As it is unlikely that disguise featured events changed the Ditto rate, we decided to combine these two datasets in the reported Ditto rate of 2.00% (95% CI [1.78, 2.25%]).

In contrast, both the April Fools’ and TCG events specifically featured Ditto in their announcements and exhibited a boosted Ditto rate during their respective events. Data from these Ditto featured events is excluded from the overall reported Ditto rate.

Lastly, we directly compared the Ditto rates for each individual featured species to their non-event counterparts. During our study, two species met their goal of 600 observations in both non-event and disguise featured event periods: Gastly, appearing in A Mega Moment, and Surskit, in the Water Festival 2022 events.

Disguise — Period Rate 95% CI
Gastly — Non-Event 1.53% [0.60, 3.25%]
Gastly — Mega Moment 2.07% [0.97, 3.93%]
Surskit — Non-Event 1.49% [0.73, 2.73%]
Surskit — Water Festival 1.76% [1.28, 2.36%]
Drowzee — Festival of Colors 2.38% [1.78, 3.12%]

We found no difference between non-event and event rates for Gastly (χ²(1) = 0.58, p = 0.30) or Surkit (χ²(1) = 0.19, p = 0.66). We did not obtain sufficient observations of Dwebble or non-event Drowzee to effectively compare their respective event and non-event Ditto rates.