What Are the Odds? Using DNA Painter's WATO Tool to Place a DNA Match

Cousin Chart Editorial · Reviewed by Joan Ramos · Updated October 2026 · 23 min read

If you've searched for a "what are the odds DNA" tool, you're probably looking for What Are the Odds?, or WATO, a free tool from DNA Painter. It helps you work out where a person fits in a family tree from how much DNA they share with several known relatives. You build a small tree, enter the shared centimorgans, mark the spots where the person might belong, and WATO gives each spot a score that shows how likely it is compared with the others.

It's one of the most useful tools for adoptees, people with an unknown parent, and anyone trying to place a puzzling DNA match. It's also easy to misuse. This guide explains what WATO does, how to build the hypothesis tree step by step, how to read the scores, the mistakes that trip people up, and a full worked example from start to finish.

What the What Are the Odds DNA tool actually does

WATO lives on DNA Painter, the genetic genealogy site built by Jonny Perl. According to DNA Painter's WATO FAQ, the tool helps you place a target into a known family tree, based on the centimorgans the target shares with people in that tree. You build a skeleton tree, enter the shared cM values, add hypotheses (possible positions for the target), and WATO scores each one. A higher score means that position is statistically more likely.

Two terms you'll see everywhere:

• Target: the person you're trying to place. Often it's you, an adoptee, or a parent or grandparent whose origins are unknown. It can also be a mystery match you want to fit into a tree.
• Hypothesis: one possible position for the target in the tree. "Pat is a grandchild of Daniel" is a hypothesis. So is "Pat is Daniel's child."

The amounts of shared DNA are measured in centimorgans (cM), a unit of genetic length. Your DNA testing company reports the total shared cM for every match. If you're new to the unit, our centimorgans explainer covers the basics.

Where WATO came from

In a June 2020 post, Upcoming webinar on WATO, including new features, Jonny Perl explains the history. Genetic genealogist Leah Larkin came up with the idea of applying probabilities to genealogy questions and worked with Andrew Millard on a spreadsheet version. That version was powerful but hard to use, so Perl built the tree-based web version in late 2017 and launched it in summer 2018. A v2 beta followed in 2020.

The WATO tool page says the shared DNA probabilities it uses were simulated by AncestryDNA and described in Ancestry's matching white paper, and that the math turning those probabilities into a score was developed by Dr. Andrew Millard. The page also notes that the tool follows Leah Larkin's "Science the heck out of your DNA" series.

What WATO is not

WATO is not a relationship predictor for a single match. If you just want to know "what could a 180 cM match be?", the Shared cM Project tool, also on DNA Painter, is the right place to start. WATO earns its keep when you have several matches who are already placed in a known tree, and you want to test where one more person fits among them.

Before you start: what you need

The WATO FAQ calls it an advanced tool and lists a few things to have in hand first. Skipping these is the number one reason people get confusing results.

1. A clear research question. For example: "Who are the parents of Pat?" or "Which of these four siblings is my great-grandfather's father?"
2. A known common ancestor. You should already have identified the most recent common ancestor (MRCA) of two or more of your matches and know exactly how each match descends from them. Tools like shared matches and the Leeds Method help you get there.
3. Shared cM values between the target and each tested person in the tree, from the testing company or a one-to-one comparison.
4. Approximate birth years for the target and for people in the tree. They aren't strictly required, but you need the target's birth year to use the "suggest hypotheses" feature.

If you're not there yet, the FAQ suggests going back to your matches' trees and the shared matches feature to sort matches into a descendant tree first. Our guide to shared DNA matches walks through that step.

How much DNA should your matches share?

The FAQ says the underlying statistics don't go below 40 cM, so most of your matches should share at least that much, and more is better. The probability data covers relationships up to the 9th degree, which includes fourth cousins, third cousins twice removed and half third cousins once removed. WATO is most accurate when most of your relationships are closer than that and most matches are above 40 cM.

How to build the hypothesis tree, step by step

The WATO tool page lists the basic steps. Here they are with practical tips added.

Step 1: Enter the target

Open WATO and enter the target's name and birth year. If you're placing yourself, you're the target. If you're helping an adoptee, they are. The birth year matters because WATO uses parental age limits to rule out impossible positions.

Step 2: Build the skeleton tree

Start from the root: the common ancestral couple you've identified. Hover over a person and click add child to build down. Alternatively, use the Load option to import a GEDCOM file and pick the couple whose descendants you want. The tool page warns that importing replaces anything already in the tree, and it offers a privacy option that swaps living people's names for initials.

What to include:

• Every tested match and the full line from the root couple down to them.
• All children of the root couple who reached adulthood, whether or not their descendants tested. The FAQ recommends this, because untested branches are exactly where the target might fit.
• Birth and death years wherever you have them. They help WATO narrow results.

You don't need every descendant. A lean tree is easier to read and faster to score.

Step 3: Enter each match's shared cM

Hover over each tested person and click Enter match cM. Type the total shared cM between that person and the target. The FAQ says to use each company's shared DNA figure for AncestryDNA, 23andMe and MyHeritage, and a default one-to-one comparison for GEDmatch. You can mix companies in one tree.

If both a parent and their child have tested and both match the target, you can include both, but the FAQ notes WATO only considers the parent. The child's DNA is just a subset of the parent's, so it adds no new information.

Step 4: Add hypotheses

Hover over a spot in the tree and click Add hypothesis. Each hypothesis is a node where the target might belong. You can also click Suggest hypotheses, and WATO will add every position it considers possible. The tool page says this needs at least one match of 40 cM or more and the target's birth year.

The FAQ's advice here is important: consider every genealogically possible position, including unknown half-siblings. Don't only test the positions you suspect. If the right answer isn't on the board, WATO can't find it. A good way to start is to put your biggest match's cM into the Shared cM Project tool and see which relationships fit.

"Suggest hypotheses" often adds unlikely options too. You can right-click and remove any that make no sense genealogically, or clear them all with the remove-suggested button.

Step 5: Read the scores

Each hypothesis now shows a score. Click any score to see the details. Below the tree, a score calculation table lists each match's cM, the relationship each hypothesis implies, and the probability of that cM for that relationship. We'll cover how to read this next.

Step 6: Save your work

The tool page explains you can save locally in your browser or to a free DNA Painter account. Local trees are lost if you clear your browser cache, so an account is safer for long projects. You can also download the tree as a file and share a link once it's saved.

How the What Are the Odds DNA scores work

This is the part most people find confusing, so let's take it slowly.

The basic idea

For each hypothesis, WATO looks at every match in turn. It asks: "If the target were here, this match would be the target's second cousin. How likely is it that second cousins share this many cM?" It does that for every match, then combines those probabilities into one figure for the hypothesis. A hypothesis where every match's cM fits comfortably gets a high combined figure. One where even a single match's cM is a poor fit gets a much lower one.

WATO then shows these as relative odds. According to the FAQ:

• A hypothesis that's impossible with your data scores 0.
• The least likely possible hypothesis scores 1.
• Every other score is compared to that one. If hypotheses score 100, 5 and 1, the top one is 100 times as likely as the lowest and 20 times as likely as the middle one.

The focus mode post on the DNA Painter blog adds one more rule: a hypothesis also scores zero if it's impossible because of birth or death dates, or if it's more than a million times less likely than the top hypothesis. With only one hypothesis on the board, the score can only be 0 or 1.

The scores are relative, not absolute

Because everything is measured against the weakest possible hypothesis, the raw number means nothing on its own. The FAQ makes this point directly: is a score of 24 or 24,000 better? Neither tells you anything alone. A score of 24,000 next to 23,900 isn't compelling at all, because the two are nearly equal.

What matters is the ratio between the top hypotheses. The FAQ gives the example of scores of 1,800, 900, 5 and 1. The 5 and 1 are probably negligible, but 1,800 and 900 are close: the top is only twice as likely as the runner-up.

Why a score of zero happens

A zero usually means at least one match's cM falls outside the range for the relationship that hypothesis implies. The FAQ says to check the Collated Match Data table, where non-viable cM values appear in red. The tool page adds that a probability shown as 0.00% in red means the relationship isn't statistically possible for that cM amount.

The FAQ also warns that one misplaced person in your tree can make a correct hypothesis score zero. Before you throw out a zero, double-check that every match is in the right spot.

Focus mode

Focus mode is a handy way to see why a hypothesis scored the way it did. Click a hypothesis's score and it gets a glow. Each match then shows a badge with its relationship to that position and the probability of its cM for that relationship. Press Escape or click the x in the blue box to leave. It's the fastest way to spot the one match that's dragging a hypothesis down.

A full worked example: placing an adoptee

Let's walk through a realistic case from scratch. The family is made up, and the probabilities in the scoring step are simple round numbers we chose to show how the math works. When you run WATO yourself, it supplies its own probabilities, which you'll see in the score calculation table. The cM ranges we compare against are real figures from the Shared cM Project.

The question

Pat was born in 1962 and adopted at birth. Pat's three strongest DNA matches all descend from the same couple, George and Ellen Hart, who married in 1905. The question is: where does Pat fit among the Harts' descendants?

The known tree

George and Ellen had four children who reached adulthood:

• Albert (born 1906). His daughter Ruth (born 1935) has tested.
• Bessie (born 1909). Her son Walter (born 1934) has died, but Walter's son Tom (born 1960) has tested.
• Clara (born 1912). Her daughter Joan (born 1936) has died, but Joan's daughter Linda (born 1958) has tested.
• Daniel (born 1915). He moved away in the 1930s. No descendants have tested, and the family knows little about him.

Pat's shared DNA: Ruth 410 cM, Tom 225 cM, Linda 200 cM.

Step 1: A first look with the Shared cM Project

Before opening WATO, check what each number could mean. Here are the relevant ranges from the Shared cM Project 4.0 tool:

Ruth's 410 cM is right next to the 1C1R average of 433. Tom's 225 and Linda's 200 are both near the second cousin average of 229. That's a strong hint that Pat is in the same generation as Tom and Linda, through a branch that isn't Albert's, Bessie's or Clara's. But other positions are possible too, and that's what WATO is for.

Step 2: Build the tree in WATO

1. Enter Pat as the target, born 1962.
2. Add George and Ellen as the root couple.
3. Add all four children: Albert, Bessie, Clara and Daniel, with birth years.
4. Add Ruth under Albert; Walter then Tom under Bessie; Joan then Linda under Clara.
5. Enter cM: Ruth 410, Tom 225, Linda 200.

Step 3: Add the hypotheses

Now list every position that's genealogically possible for someone born in 1962. We'll test five:

• H1: Pat is a grandchild of Daniel (an unknown child of Daniel had Pat).
• H2: Pat is a grandchild of a half-sibling of the four Hart children, say a child George had with another partner.
• H3: Pat is Daniel's own child. Daniel would have been 47 in 1962, which is within WATO's age limits for a father.
• H4: Pat is a grandchild of Clara, through a child other than Joan.
• H5: Pat is a grandchild of Albert, through a sibling of Ruth.

Step 4: Work out the relationships each hypothesis implies

Right away, WATO would score H4 and H5 as 0. In the score calculation table, Linda's value under H4 and Ruth's value under H5 would show in red.

Step 5: See how the scores come out

For the three possible hypotheses, WATO looks up the probability of each match's cM for each implied relationship and combines them. To show the mechanics, here are round illustrative probabilities (WATO's real values will differ):

• H1: Ruth as 1C1R at 410 cM: 0.30. Tom as 2C at 225 cM: 0.40. Linda as 2C at 200 cM: 0.40. Combined: 0.30 x 0.40 x 0.40 = 0.048.
• H2: Ruth as half 1C1R: 0.10. Tom as half 2C: 0.15. Linda as half 2C: 0.20. Combined: 0.10 x 0.15 x 0.20 = 0.003.
• H3: Ruth as 1C: 0.02. Tom as 1C1R: 0.25. Linda as 1C1R: 0.20. Combined: 0.02 x 0.25 x 0.20 = 0.001.

Now divide each by the smallest possible one (H3, 0.001) to get relative scores:

• H1: 0.048 / 0.001 = 48
• H2: 0.003 / 0.001 = 3
• H3: 0.001 / 0.001 = 1
• H4 and H5: 0

Step 6: Interpret the result

H1 leads. It's 48 times as likely as H3 and 16 times as likely as H2 (48 divided by 3). That's a meaningful lead, and it points to Daniel's branch. But H2 isn't dead. A 16-to-1 edge can still be wrong, and H2 represents a scenario (an unknown half-family) that genealogists see often in real cases.

Remember the FAQ's warning: "WATO uses probabilities to steer you in the most likely direction, but cannot provide proof on its own." So the result tells Pat where to research next, not who Pat's grandparents are.

Step 7: Add a match that separates the top two

What would split H1 from H2? Under H1, Pat descends from both George and Ellen. Under H2, Pat descends from George only. So a match with Ellen's side of the family, someone who descends from Ellen's sister, for example, would be a third cousin or closer under H1 and no relation at all under H2.

Suppose Pat finds a 95 cM match with Ellen's sister's great-grandson and confirms his tree. Under H1 he'd be Pat's third cousin, since both descend from Ellen's parents through four generations, and 95 cM sits comfortably in the Shared cM Project's third cousin range (average 73 cM, range 0 to 234). Adding Ellen's parents and that line to the WATO tree would sharply favor H1, because under H2 Pat shouldn't share any DNA with Ellen's family at all. If instead Pat shares nothing with several tested descendants of Ellen's siblings, H2 gets stronger. The FAQ notes that confirmed zero matches carry weight too.

Step 8: Turn to the records

With Daniel's branch as the lead, the next step is records: Daniel's marriage, his children's births, obituaries and city directories for wherever he moved. If Daniel had a child who was the right age to be Pat's parent in 1962, that becomes the next target for targeted DNA testing. Our guide to unknown parent DNA searches covers that stage, and our article on how to confirm a DNA match explains what counts as solid confirmation.

Common mistakes with the What Are the Odds tool and how to avoid them

1. Leaving the right answer off the board

WATO only compares the hypotheses you give it. If Pat's real position is in a branch you never added, WATO will still crown a winner, just the wrong one. Always add all children of the root couple and consider half-sibling lines. "Suggest hypotheses" helps here.

2. Entering a misplaced match

A match in the wrong spot can push the right hypothesis to zero. The FAQ is direct about this. Before trusting a result, double-check each match's tree, especially for matches whose trees you built yourself from records.

3. Counting the same DNA twice

Adding a parent and child, or several siblings, as separate data points makes the tool overconfident. The FAQ says the calculations become overconfident when you include several closely related people, and suggests testing the sensitivity by keeping just one sibling. If the ranking holds up with one sibling, you can trust it more. If it flips, your data isn't as strong as it looked.

4. Entering your own close relatives' cM with you

If you're the target, you can add your brother or a known cousin to the tree as a visual reminder, but the FAQ says not to enter their shared DNA with you, because you already know that relationship.

5. Using WATO in an endogamous family

The FAQ says WATO isn't meant for endogamous populations, where people are related through many lines at once, and that there isn't good probability data to model it. Endogamy raises cM values across the board, so hypotheses that are too close often score higher than they should. If your family comes from a community with a lot of intermarriage, read our guide to endogamy first and treat WATO results with extra caution.

6. Ignoring recent pedigree collapse

If, say, the target's grandparents were first cousins, the target shares extra DNA with some relatives. The FAQ says recent pedigree collapse can lead WATO to wrongly rule out possible positions, while more distant collapse usually matters less. See pedigree collapse explained for how it works.

7. Relying on small matches

Matches under 40 cM are below the range of WATO's statistics. They can also come from a completely different, older line. Use them as supporting clues, not as the backbone of a WATO tree.

8. Forgetting about ages

The FAQ lists WATO's age assumptions: mothers between 12 and 55, fathers between 12 and 75, and half-siblings no more than 30 years apart. Enter birth years so the tool can rule out impossible positions. If a hypothesis survives only because you left dates blank, it may be wasting a slot.

9. Treating a high score as proof

Both the FAQ and the focus mode post say it plainly: a high score doesn't prove a hypothesis is right. It tells you where to look. Proof comes from records, more DNA testing, and agreement between the two.

10. Reading raw scores instead of ratios

A score of 5,000 sounds impressive. If the runner-up is 4,000, it isn't. Always compare the top score with the second-highest.

Special cases: siblings, descendants and mystery matches

Two siblings matching the same person

The FAQ describes an alternative to adding both siblings separately. Make one sibling the hypothesis and enter each sibling's cM with the match, labeled clearly (for example "Diane-Bob" and "Diane-Ken"). That captures both siblings' DNA without double-counting the tree position.

Your sister gets a different answer

That's normal. Siblings inherit different pieces of DNA from each grandparent, so their cM values with the same matches differ. The FAQ suggests focusing on the parts of the tree that rank highly for both of you. Our page on half siblings vs. full siblings explains how much sibling DNA usually varies.

Placing an ancestor who never tested

Say you want to find your maternal grandfather's parents. He never tested, but you did. In that case, you're the target. The FAQ explains how to repeat the path from the unknown ancestor down to you in each possible position, marking each copy as a hypothesis. If several of his descendants tested, only one can be the hypothesis at a time. Copy the tree for each tested person and compare.

A mystery match who tested at another company

WATO needs the target's cM with each person in the tree. The FAQ notes you can get those figures from MyHeritage's shared matches, 23andMe's DNA comparison feature, or a GEDmatch one-to-one comparison using the match's kit number.

A confirmed non-match

Sharing no DNA with someone is information. Per the FAQ, a confirmed zero can rule out relationships such as second cousin or closer. Distant cousins often share nothing, though. In How many genomic blocks do you share with a cousin?, Graham Coop's lab at UC Davis estimates that roughly 25% of fourth cousins share no DNA through their common ancestors, so a zero with a fourth cousin proves little.

WATO compared with other tools for placing a match

WATO works best alongside other methods, not instead of them. Here's how it fits with the tools most genealogists use.

Why ranges overlap, and why that's the reason WATO helps

Look back at the Shared cM table in the worked example. A 225 cM match could be a second cousin (range 41 to 592), a 1C1R (102 to 980), a half 1C1R (62 to 469), or a half second cousin (10 to 325). One match alone can't tell these apart.

This overlap is real, not a flaw in the data. Blaine Bettinger's Shared cM Project Version 4.0 report drew on nearly 60,000 submitted relationships, and it shows wide, overlapping ranges for many relationship groups. The report also lists the main sources of error in crowd-sourced data, including mistaken relationships, misunderstood "removed" relationships, endogamy, pedigree collapse, and differences in company matching thresholds.

WATO's advantage is that it combines several matches at once. Each match narrows the options a little. Three or four well-placed matches in different branches can separate hypotheses that no single match could. That's the whole point of the tool, and it's why adding a match from a different branch is usually the best way to break a tie.

For a full table of ranges, see our shared cM chart.

Wrapping up: using WATO well

WATO turns a pile of cM numbers into a ranked list of where a person probably fits in a family tree. Used well, it saves months of guesswork. Keep these points in mind:

• Start with a clear question and a known common ancestor.
• Add all children of the root couple and every reasonable hypothesis, including half-relationships.
• Use matches of 40 cM or more, and enter birth years.
• Read the ratio between the top scores, not the raw numbers.
• Check zeros for misplaced matches before discarding them.
• Treat the result as a research lead, then confirm it with records and more testing.

Once you've pinned down where your match fits, you'll want the right name for the relationship. Our free family relationship calculator gives you the exact term, from second cousin once removed to half first cousin.

Keep reading: Coefficient of Relationship: How to Work Out How Related Two People Are.

Frequently Asked Questions

Is DNA Painter's What Are the Odds tool free?

Yes. WATO is free to use on DNA Painter, and you don't need an account to build and score a tree. The tool page notes that you can save trees locally in your browser, but those are deleted if you clear your browser cache. Signing up for a free DNA Painter account lets you save trees to your account and share them with a link. That's the safer choice for any project you expect to work on over weeks or months.

How accurate is WATO?

WATO is only as accurate as the tree and cM values you give it. DNA Painter describes the calculations as approximate and warns they become overconfident when several closely related people are included. The underlying probabilities cover relationships up to the 9th degree and down to 40 cM, so results are most reliable when most matches are closer than fourth cousins and share more than 40 cM. Endogamy and recent pedigree collapse can distort scores. Treat WATO as a way to rank possibilities, then confirm with records.

What is a good score in What Are the Odds?

There's no single good score, because WATO scores are relative. The least likely possible hypothesis is set to 1, and everything else is compared to it. What matters is the gap between the top two hypotheses. A top score of 1,800 next to 900 means the leader is only twice as likely. DNA Painter's FAQ says that if your highest score is under 20, your data doesn't distinguish well between the options, and you should add more matches.

Why does my hypothesis score zero in WATO?

A zero means the hypothesis isn't possible with your data. Usually at least one match's cM falls outside the range for the relationship that position would create, which shows in red in the collated match data or score calculation table. A hypothesis can also score zero because of birth or death dates, or if it's more than a million times less likely than the top hypothesis. Check for a misplaced match or a typo before you discard it, because one error can zero out the right answer.

How many DNA matches do I need to use WATO?

There's no fixed number, but more well-placed matches give better results. With only one match, WATO can't do much more than the Shared cM Project tool. Three or more matches who descend from different children of the root couple work best, because each one tests the hypotheses from a different angle. Avoid stacking up siblings or parent-and-child pairs, since they repeat the same information and make the scores overconfident. Matches should mostly be above 40 cM.

Can WATO find my biological father?

WATO can help narrow down where your biological father fits in a family, but it can't name him on its own. You first need matches whose common ancestors you've identified, so you can build a tree of that family. WATO then ranks the possible positions for you or your father. The top-ranked spots tell you which branch and generation to research. From there, records, family contact and targeted DNA tests of likely relatives are what confirm a specific person.

What's the difference between WATO and the Shared cM Project tool?

The Shared cM Project tool looks at one match at a time and lists relationships that fit the cM you enter, based on Blaine Bettinger's crowdsourced data. WATO works with a whole tree of matches at once. It tests specific positions for one target and scores them by how well all the matches' cM values fit together. Use the Shared cM tool to explore a single match, and WATO when you have several placed matches and a defined question.

Does WATO work with Ancestry, 23andMe and MyHeritage matches?

Yes. DNA Painter's FAQ says you can mix testing companies in one WATO tree. For AncestryDNA, 23andMe and MyHeritage, use the shared DNA amount each company reports. For GEDmatch, use a one-to-one comparison with the default settings. If a target tested at 23andMe, MyHeritage or GEDmatch, you can often get their shared cM with other people through shared match views, the DNA comparison feature, or a one-to-one comparison with a kit number.

Sources

We checked every source below while writing this guide. Last reviewed October 30, 2026, by Joan Ramos.

1. What Are the Odds? (WATO) tool (DNA Painter)
2. Frequently Asked Questions about WATO (DNA Painter)
3. Focus mode in What are the Odds? (WATO) (DNA Painter Blog (Jonny Perl))
4. Upcoming webinar on WATO, including new features (DNA Painter Blog (Jonny Perl))
5. The Shared cM Project 4.0 tool (DNA Painter (data by Blaine T. Bettinger))
6. The Shared cM Project, Version 4.0 (March 2020) (Blaine T. Bettinger, The Genetic Genealogist)
7. How many genomic blocks do you share with a cousin? (Coop Lab, University of California, Davis (gcbias blog))