It is common practice for some end-users (or re-sellers) of large quantities of the same part number to test and "bin" all received parts according to the results of the test. With automated test equipment, it is not necessary to just sample a lot of parts to make an accept/reject decision of the whole lot, which was a common procedure in the previous century. While statistically any device from a given production run will have characteristics that fall between the published or "guaranteed" minimum and maximum values, with "most" of the parts having characteristics close to the typical value, the actual range of values depends on how well the production process is understood and controlled.
Unless there are specific quality control documents and specifications that all parts must adhere to, the only remedy an end-user has is to return to the manufacturer or the re-seller any parts that do not fall within the limits of the guaranteed values. Some purchase orders may specify that an entire lot will be returned if any of the parts fails to meet specifications. Obviously such practices will inflate the cost of the part. If the end-user has a very large volume commitment, they may bin the parts after testing and sell the outliers at a discount, reserving for themselves those bins closest to the "nominal" value. These practices clearly put the hobbyist buying in onesie-twosie quantities at a disadvantage, but there is little that can be done about it. The Golden Rule applies: He who has the Gold, Rules.
Somewhat disgusted, not mad, that you have failed to understand. Where did you find the value of 6.8 pF? What do you mean to "divide 6.8 pf with 2V and 20V? If you are going to divide anything, pick any number between the minimum capacitance ratio of 5.0 and the maximum capacitance ratio of 6.5 and divide that capacitance ratio number into any value between the minimum and maximum capacitance values that occur with 2 V reverse bias. This range of values represents the maximum varactor capacitance. The result after dividing by the capacitance ratio will be the minimum varactor capacitance.
The minimum varactor capacitance occurs with 20 V DC reverse bias. The maximum varactor capacitance occurs with 2 V DC reverse bias. The varactor will have more capacitance for reverse bias voltages less than 2 V, but the capacitance is not specified nor is it guaranteed. The varactor will have less capacitance for reverse bias voltage more than 20 V, but the capacitance in not specified nor is it guaranteed. Exceeding 25 V reverse bias may irrevocably damage the varactor. It is inadvisable to forward-bias the varactor diode without providing a means to limit the forward current, but it is your part, you own it: do what whatever you feel is necessary to advance your learning process.