Background

Questioninfo icon

Who Killed Tair Rada?

19 Jan, 2023

Answerinfo icon

Ola Kravchenko murdered Tair Rada.

(98% probability)

Backgroundinfo icon

On 6th December 2006, Tair Rada, a 13-year-old Israeli schoolgirl, was found dead, covered in her own blood and her throat sliced, inside a locked bathroom stall in her school, Nofey Golan, in the town of Katzrin.

Initial suspects were Roman Zdorov, a floorer temporarily working at the school and a new immigrant to Israel from Ukraine, as well as Tair’s schoolmates. During the police investigation, Zdorov confessed to the crime but recanted soon after. Zdorov was convicted of the murder on 14th September 2010.

In 2012, Adir Habani told the police that his ex-girlfriend, Ola Kravchenko, had confessed to him on the day of the murder that she had killed Tair, and shown him bloodied clothes and a knife.

Adir’s testimony became public in 2016 with the international release of the documentary series Shadow of Truth and has continued to receive international attention with a true crime podcast in 2021.

Following the emergence of new evidence, and awaiting a rare retrial ordered by the Supreme Court, Zdorov was released on house arrest on 26th August 2021.

During the retrial, which started on 12th October 2021, an old piece of evidence, a hair found at the crime scene, was found to strongly match Adir Habani. This evidence dramatically increases the likelihood that Ola murdered Tair since Habani has an alibi and he claims that Ola wore his clothes while committing the murder. However, this analysis intentionally ignores this evidence, in order to demonstrate what a probabilistic analysis could have found using only the evidence available in 2012.

Hypotheses Consideredinfo icon

Calculated Resultsinfo icon

Calculated Resultsinfo icon

1

98.4%
Ola:

Ola Kravchenko murdered Tair Rada.

98.4%

2

1.3%
Zdorov:

Roman Zdorov murdered Tair Rada.

1.3%

3

0.4%
Schoolmates:

Between 1-4 Nofey Golan students murdered Tair Rada.

0.4%

Starting Pointinfo icon

Initial Probabilities

Name

Updated Likelihoods
Zdorov
0.2%
Schoolmates
7%
Ola
93%

As no clear motive exists for Zdorov or the Schoolmates, and no statistics exist for the unique circumstances of Tair’s murder in a school, in order to arrive at initial likelihoods for each of these hypotheses, Rootclaim looked at statistics on the identity of school shooters in the U.S. and assigned these to the schoolmates and Zdorov hypotheses. We then increased Zdorov’s odds of murdering due to being male and a temporary worker, and adjusted according to how many students were studying at Nofey Golan at the time. As Ola is known to harbor a strong urge to kill random people, her hypothesis was assessed based on the specific likelihood of her wanting and being able to murder Tair. Given the rarity of school murders in Israel, the coincidence of one happening in proximity to Ola puts her as the most likely murderer, prior to examining the specific evidence of the case.

Name

Updated Likelihoods
Zdorov
0.2%
Schoolmates
7%
Ola
93%

Evidenceinfo icon

Effectinfo icon

Zdorov's behavioral profile

Name

Effectinfo icon

Updated Likelihoods

Zdorov
×3
0.2%
Schoolmates
×1
2.3%
Ola
×1
97%

Zdorov’s investigation and the search of his belongings resulted in findings indicating behaviors that increase the likelihood of his guilt: he kept a knife collection, read an article on how to neutralize an enemy with a knife, and claimed to have thrown away the pants he was wearing the day of the murder. 

Yet, just as a bigger fishing net catching more fish is not indicative of more fish in the area, the wider and more thorough the search and investigation, the less weight should be attributed to each piece of evidence arising from it. Thus, as Zdorov’s physical and digital life was combed through, we would actually expect to see evidence of similar level, even from someone innocent. As such, Zdorov’s behavior as a whole only slightly raises the likelihood of his guilt.

 

Name

Effectinfo icon

Updated Likelihoods

Zdorov
×3
0.2%
Schoolmates
×1
2.3%
Ola
×1
97%

Schoolmate's behavioral profile

Name

Effectinfo icon

Updated Likelihoods

Zdorov
×1
0.2%
Schoolmates
÷4.3
0.5%
Ola
×1
99%

Although some minor claims of strange behavior or lying have been made against them, Tair’s schoolmates did not exhibit any unexpected behaviors after Tair’s murder. Moreover, despite 10 or more students being interrogated, no confessions were made nor were any substantially incriminating statements or evidence given to the interrogators. Neither were such statements made in the years since the murder. Studies, and our own analysis of teen murderers, found children to be likely to confess under investigation, as well as spontaneously confess to friends. The lack of confessions or suspicious activity by the schoolmates, significantly reduces the likelihood for their hypothesis. 

 

Name

Effectinfo icon

Updated Likelihoods

Zdorov
×1
0.2%
Schoolmates
÷4.3
0.5%
Ola
×1
99%

Zdorov's Confessions

Name

Effectinfo icon

Updated Likelihoods

Zdorov
÷1.2
0.9%
Schoolmates
÷6
0.5%
Ola
÷6
99%

After Zdorov’s arrest, he was interrogated intensively by police investigators, and a Russian-speaking undercover informant, Arthur, was planted in his prison cell with the promise of financial rewards if he could elicit a confession from Zdorov.

Zdorov eventually confessed and reenacted the murder on 19th December 2006. A day later, he recanted his confession, claiming that he was under extreme pressure and that the argument between him and Tair that he claimed led to him murdering her never occurred. The weakness of his confession and his overall pattern of behavior are consistent with a false confession given in order to obtain a lighter sentence, which reduces the likelihood of his guilt. 

Zdorov’s confession has elements that may be claimed to make it more genuine, and others that make it less so. We therefore used general statistics of false confessions, which resulted in a substantial increase of the Zdorov hypothesis.

 

Name

Effectinfo icon

Updated Likelihoods

Zdorov
÷1.2
0.9%
Schoolmates
÷6
0.5%
Ola
÷6
99%

Adir's Testimony

Name

Effectinfo icon

Updated Likelihoods

Zdorov
÷444
0.1%
Schoolmates
÷444
0.05%
Ola
÷5
100%

Adir Habani, Ola’s ex-boyfriend, testified to the police in 2012 that Ola had confessed to him about murdering Tair the day the murder was committed. He recounted Ola’s story of how she entered the school as well as showing him the bag in which she kept her blood-stained clothes and knife. The police dismissed this at the time as the ramblings of a vengeful ex. However, the testimony’s high level of accuracy, the rarity of false accusations of a confession, and two further claims of Ola confessing, increase the likelihood that Adir’s testimony is true. Overall, this substantially increases the likelihood of the Ola hypothesis. 

 

Name

Effectinfo icon

Updated Likelihoods

Zdorov
÷444
0.1%
Schoolmates
÷444
0.05%
Ola
÷5
100%

Peace Symbol Engraving

Name

Effectinfo icon

Updated Likelihoods

Zdorov
÷1587
0.1%
Schoolmates
÷1587
0.05%
Ola
÷171
100%

A peace symbol was carved onto the door of the toilet stall in which Tair was murdered. The symbol has been found in Ola’s jewelry and art around the time of the murder, slightly increasing the likelihood that Ola is the murderer.

Name

Effectinfo icon

Updated Likelihoods

Zdorov
÷1587
0.1%
Schoolmates
÷1587
0.05%
Ola
÷171
100%

No forensic evidence incriminating Zdorov and Ola

Name

Effectinfo icon

Updated Likelihoods

Zdorov
÷1.5
0.07%
Schoolmates
×1
0.06%
Ola
÷1.1
100%

No forensic evidence was found that linked Zdorov to Tair’s body or to the crime scene. Similarly, none of Tair’s DNA was found on Zdorov’s belongings, despite the extraordinary effort to find her DNA. As such, this reduces the likelihood of the Zdorov hypothesis.

The crime scene similarly did not include forensic evidence of Ola (other than Adir’s hair which is purposefully ignored). However, since she is claimed to have come prepared, and was not thoroughly investigated, the Ola hypothesis is reduced by a smaller factor.

Name

Effectinfo icon

Updated Likelihoods

Zdorov
÷1.5
0.07%
Schoolmates
×1
0.06%
Ola
÷1.1
100%

Ignored Evidence

Discussioninfo icon

userIcon
user icon
Ollie4943
Jan 19, 2023 at 8:36 AM
1. Your model is too conservative. This claim is backed only by my intuition and not by any concrete calculations. I believe that based on the evidence you presented, the Ola hypothesis is even more likely than the 98% you are giving it 2. Not all facts were considered. More evidence against Ola was presented by Zdorov’s attorney. Furthermore, your model should also take into consideration facts that are not admissible in the court of law (e.g., Ola’s refusal to undergo a polygraph test). 3. The likelihood of Zdorov committing the murder is too high in comparison to the other non-Ola possible culprits. Overall, great stuff. Further analysis that takes into account the hairs that were found in the crime scene would be appreciated.
user icon
Truth seeker
Apr 1, 2023 at 11:05 AM
Absolutely Brilliant! Please PLEASE reach Ilana Rada and provide her with the data, as it could help her in the ongoing case (that is now most likely to re-open) and give her and Tair justice and closure. I truly believe that if this data is presented in court in a clear and thorough manner, it could affect the verdict for the best.
logo icon
Rootclaim
Jan 23, 2023 at 1:47 PM
Thank you for your comment Ollie. We agree, the likelihood of the Ola hypothesis is likely higher. However, our methodology assigns ranges to each number which serve as a safety margin to account for possible errors and uncertainties on our part. Using ranges tends to result in the weakening of the leading hypothesis and the strengthening of others. That Ola’s hypothesis reaches 98%, and that Zdorov’s is as low as 1.3% even with this room for error, speaks to the strength of the Ola hypothesis - it's one of our most confident conclusions yet. As for missing facts, we take into consideration the most important pieces of evidence which we believe to have the strongest effect. Once a preponderance of evidence is reached, the smaller pieces of evidence have too small an effect to have any real bearing on the conclusion. However, as mentioned in the introduction, specifically in this analysis we purposefully ignore one piece of evidence which has a strong effect: the hair which is a strong match with Adir Habani. We did this to show that even before forensic evidence was available, it would have been exceedingly hard to convict Zdorov. We suspect this evidence would boost the Ola hypothesis to 99%+.