What those scholars who have counted the number of authentic hadiths meant when they said ‘authentic’, is in terms of the science of rijal and they had no intention to narrow things down in fiqhi and intellectual discussions, because in the view of those same scholars of the past and today, the criterion for the authenticity of a hadith isn't limited only to the reliability of the narrator, rather, the criteria for the authenticity and acceptability of a hadith is broader than that; anything that brings us to certitude regarding the authenticity of a hadith is what matters, not only the reliability of the narrator. This is in accordance with the methodology of past scholars; they would call any hadith they were sure had been issued by an infallible to be authentic, meaning that after assessing the chain and content of a hadith, if they became sure that it is from an infallible, they would deem it authentic and categorize it as ‘sahih’.
In order to see to what extent the given statistics are accurate, we must pay attention to several points:
A) Early scholars would divide hadiths into the two categories of authentic (sahih) and inauthentic. What they meant by authentic or sahih hadiths was any hadith that had inner clues and evidence (e.g., the reliability of the narrator) or outer clues (e.g., the hadith being in a reliable asl) proving its authenticity, making a hadith that lacked the two, inauthentic in their eyes. However, from the time of Seyed Ahmad bin Tawus, this method was changed and hadiths were divided into four groups; one group were inauthentic, and the other three were subcategories of authentic ones, named: sahih, hasan and muwathaq.
1- If all of the narrators of the hadith are Imamite, praised and just, the hadith will be sahih.
2- If a narrator is an Imamite and praised, but there is no explicit mention of his being just, that narrator will be hasan, hence the hadith is called hasan.
3- If the narrator(s) of the hadith are reliable, but aren't Imamites, that hadith will be muwathaq.
This quadruple categorization was first done in the seventh century, but Kuleini, compiler of Kafi, has gathered all of the sahih hadiths in the terms of early scholars. In his view, these hadiths can be authentic for a number of reasons, such as there being evidence proving their authenticity, or the hadith being in several asls of the Usul Arba’mi’ah, or the hadith having numerous chains of narrators, or the hadith being in an asl which has an author whose reliability is consensual, etc.
B) In the view of contemporary scholars, the criteria for the authenticity of a hadith, lays in the certitude one has regarding the hadith, and isn't restricted only to the narrator being reliable. This also reads with what the early scholars deemed authentic; any hadith that they were sure had actually been issued by an infallible. Meaning that if after assessing the chain of narrators and the content of a hadith, they came to the conclusion that it had definitely been issued by an infallible, it was sahih, regardless of whether the chain was made up of Imamite, just and praised narrators. At the same time, those scholars who have taken on the task of counting the number of ‘sahih’ hadiths, have gone by the criterion of the science of rijal, which only looks to find the hadiths that have a chain comprised of Imamite, just and praised narrators, nevertheless, they aren't tightening the scope of fiqhi or intellectual discussions through this method.
It is because of this that Sheikh Hasan bin Zayd al-Din, who has been quoted saying that the sahih hadiths in Kafi are one fifth of the total number of hadiths therein, has said elsewhere: “ان احادیث کتبنا المعتمدة محفوفة بالقرائن و ان المتقدمین الی زمن العلامة کانوا یعملون بالقرائن لا بهذا المصطلح المشهور بعده و ان المتأخرین قد یعملون بذلک ایضاً” (“Indeed the hadiths found in our reliable books are surrounded by clues and evidence [that proves their authenticity], and the early scholars till the time of Allamah would act upon these clues, not this common term [of sahih] that was coined after him; later scholars also go by this criterion”).
C) The main problem with what has been said is the false conclusion that has been come to based on it. What we mean is that both what was said about most of the hadiths in Kafi being inauthentic isn't correct, and also the claim that the imams didn’t possess superhuman powers is also incorrect, because first of all, the reasons behind many beliefs are intellectual arguments and reasoning; hadiths are only mentioned as corroboration of what has been understood through argument, and secondly, because many of these beliefs have been given to us by series of hadiths with different content that somewhat verify each other. For example, there are many hadiths that share the same content of hadiths that have been mentioned in the two books of Bidayah al-Ma’arif al-Ilahiyyah and Rahnama Shenasi, such as: hadiths that tell us the imams are muhaddethun (hearers of words from the unseen) or that their knowledge is from the divine.
 Sobhani, Jafar, Kulliyyat fi Ilm al-Rijal, Nashr Eslami Institute, pp. 357-359.
 Sobhani, Jafar, Usul al-Hadith wa Ahkamuh, Nashr Eslami Institute, pg. 50.
 Fayd Kashani, Wafi, pp. 11-12.
 Seyfi Mazandarani, Ali Akbar, Miqyas al-Riwayah fi Ilm al-Dirayah, pg. 44. For further information, refer to Question 1937 (website: 2185).
 Wasa’el al-Shia, vol. 20, pg. 71.
 Kuleini, Kafi, vol. 1, pg. 532.
 Majlisi, Mir’at al-Uqul, vol. 3, pg. 54.