Lack of type inheritance using a backward chaining reasoner?

Question

I'm learning about reasoning and knowledge engineering and I made up the following example:

@prefix : <http://example.org#> .
@prefix owl: <http://www.w3.org/2002/07/owl#> .
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
@prefix xml: <http://www.w3.org/XML/1998/namespace> .
@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
@base <http://www.w3.org/2002/07/owl> .

[ rdf:type owl:Ontology
 ] .

:value rdf:type owl:DatatypeProperty ;
   rdfs:range xsd:string .

:humiditySensor rdf:type owl:Class ;
            rdfs:subClassOf :sensor .

:location rdf:type owl:Class .

:room rdf:type owl:Class ;
  rdfs:subClassOf :location .

:sensor rdf:type owl:Class.

:temperatureSensor rdf:type owl:Class ;
               rdfs:subClassOf :sensor .

:temp1 rdf:type owl:NamedIndividual ,
            :temperatureSensor ;
    :room 201 ;
   :value 29 .

I'm trying to infer that temp1 is of type sensor. When I pose the following query to Eye reasoner (backward chaining) I get nothing, i.e. the fact that temp1 is of type sensor is not inferred, even though subClassOf is transitive:

@prefix : <http://example.org#>.
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#>.

{
    ?x a :sensor.
    ?x :value ?y.
}
=>
{
    ?x :value ?y.
}.

If I write a couple of rules manually to express this transitivity property of subClassOf, it will work. However, CWM reasoner (forward chaining) in the proof states correctly that temp1 is of type sensor, without the additional rules.

Is this because of the difference between forward and backward chaining?

I hope I'm making sense.