MathOverflow will be down for maintenance for approximately 3 hours, starting Monday evening (06/24/2013) at approximately 9:00 PM Eastern time (UTC-4).
show/hide this revision's text 6 to make the statement correct, change of the composition, the spirit remains the same otherwise.

Many theorems have the form : Premise(es) implies Conclusion(s)

Example A of wrongness:
There are many examples in which a theorem is stated without mentioning that part of the premise is not necessary to reach the conclusion.
Usually it is simple (and much better) to add a remark stating that the result is not sharp (ideally providing an example of weaker premise holding with the solution).

But there is another type of bias :

Added Note: Below composition means the AND of two relations ( for classical composition the transitivity does not compose! ( thanks to HenrikRüping remark).

Example B of wrongness:
Theorem 1 : The composition of 2 equivalence relations is an equivalence relation.
Or in fewer words : Equivalence relations are stable under composition.

Actually there is a much finer version of B :

Theorem A: For relations each of the following properties are stable under composition : Reflexive , Transitive , Symmetric.
By conjunction of the above we obtain:
Corollary B: Equivalence relations are stable under composition

Note: The second form is not only more precise but it also makes the mention "left as an easy exercise" more acceptable.

The "WRONG" notion:
I called theorem 1 (or its statement) wrong as it induced the reader to think that the conjunction of the 3 properties plays a role in proving the conclusion.

Of course only true theorems may be qualified as wrong.

Taking an absolute stance you may call wrong any theorem that is not a tautology.
A less absolute stance would call wrong any theorem that is not a tautology and in which you forget to mention non-sharpness.

Question 1: is there a better / more adequate term than wrong ( the subtext is: do you think it is a good notion?) .

Question 2: Do you know examples that follow a pattern like B or some variation in lack of tautology?

ADDED TO BE MORE SPECIFIC:

Question 3: More specifically : Are there other types of patterns showing a distance between premise and conclusion. The types need to be common in the mathematical literature, not purely logical types ( of course those are more countable).

show/hide this revision's text 5 change antisymetric to symetric

Many theorems have the form : Premise(es) implies Conclusion(s)

Example A of wrongness:
There are many examples in which a theorem is stated without mentioning that part of the premise is not necessary to reach the conclusion.
Usually it is simple (and much better) to add a remark stating that the result is not sharp (ideally providing an example of weaker premise holding with the solution).

But there is another type of bias :

Example B of wrongness:
Theorem 1 : The composition of 2 equivalence relations is an equivalence relation.
Or in fewer words : Equivalence relations are stable under composition.

Actually there is a much finer version of B :

Theorem A: For relations each of the following properties are stable under composition : Reflexive , Transitive , AntisymmetricSymmetric.
By conjunction of the above we obtain:
Corollary B: Equivalence relations are stable under composition

Note: The second form is not only more precise but it also makes the mention "left as an easy exercise" more acceptable.

The "WRONG" notion:
I called theorem 1 (or its statement) wrong as it induced the reader to think that the conjunction of the 3 properties plays a role in proving the conclusion.

Of course only true theorems may be qualified as wrong.

Taking an absolute stance you may call wrong any theorem that is not a tautology.
A less absolute stance would call wrong any theorem that is not a tautology and in which you forget to mention non-sharpness.

Question 1: is there a better / more adequate term than wrong ( the subtext is: do you think it is a good notion?) .

Question 2: Do you know examples that follow a pattern like B or some variation in lack of tautology?

ADDED TO BE MORE SPECIFIC:

Question 3: More specifically : Are there other types of patterns showing a distance between premise and conclusion. The types need to be common in the mathematical literature, not purely logical types ( of course those are more countable).

show/hide this revision's text 4 spelling only

Many theorems have the form : Premise(es) implies Conclusion(s)

Example A of wrongness:
There are many examples in which a theorem is stated without mentioning that part of the premise is not necessary to reach the conclusion.
Usually it is simple (and much better) to add a remark stating that the result is not sharp (ideally providing an example of weaker premise holding with the solution).

But there is another type of bias :

Example B of wrongness:
Theorem 1 : The composition of 2 equivalence relations is an equivalence relation.
Or in fewer words : Equivalence relations are stable under composition.

Actually there is a much finer version of B :

Theorem A: For relations each of the following properties are stable under composition : Reflexive , Transitive , Antisymmetric.
By conjunction of the above we obtain:
Corollary B: Equivalence relations are stable under composition

Note: The second form is not only more precise but it also makes the mention "left as an easy exercise" more acceptable.

The "WRONG" notion:
I called theorem 1 (or its statement) wrong as it induced the reader to think that the conjunction of the 3 properties plays a role in proving the conclusion.

Of course only true theorems may be qualified as wrong.

Taking an absolute stance you may call wrong any theorem that is not a tautology.
A less absolute stance would call wrong any theorem that is not a tautology and in which you forget to mention non-sharpness.

Question 1: is there a better / more adequate term than wrong ( the subtext is: do you think it is a good notion?) .

Question 2: Do you know examples that follow a pattern like B or some variation in lack of tautology?

ADDED TO BE MORE SPECIFIC:

Question 3: More specifically : Are there other types of patterns showing a distance between premise an and conclusion. The types need to be common in the mathematical literature, not purely logical types ( of course those are more countable).

show/hide this revision's text 3 Added a question to be more specific ( metas/judges could remove question 2)
show/hide this revision's text 2 add a tag
show/hide this revision's text 1