Declarer never needs to count all 52 cards, just like declarer never needs to make the contract. To be consistently successful declarer simply needs to be counting more cards than defenders are counting. Of course if the defenders are counting at least 26 each, declarer should be counting a full 52.
There are two fundamental reasons why one must aspire to count all 52 cards:
- You get expanding returns, not diminishing returns, as every card you count increases the ability to locate and count the next card. The reason for this is that each located card reduces the number of cards remaining to be counted.
- All of the advanced play techniques (single-, double-, and repeating-squeezes; throw-ins, etc.) require an exact, or nearly exact, count. One cannot aspire to perform any of these reliably without an exact count; neither can one effectively defend against them without an exact or nearly exact count.
The gain in play by having a solid familiarity of advanced techniques, can approach a trick over the play of a good intermediate player who has not studied these techniques. That is like being dealt 2 or 3 extra points every hand. If your partner has likewise studied these, (s)he also gets these extra points.
Imagine how many more successful contracts you will bid, and how many more opposing contracts you will set, if you and your partner are each dealt an extra queen each hand. That is the advantage one obtains by first learning to get an exact count early in the hand, and then studying the advanced techniques that allow you to take full advantage.
This analysis may also bear o your question here concerning cycles of over- and under-bidding. If one considers the very aggressive bidding of the Meckwell partnership (Eric Rodwell & Jeff Meckstroth), it is their excellent declarer and defender play, combined with superb table feel and intuitive hand evaluation, that allows them to be successful playing this way. Barry Crane played a similarly aggressive style in his day.
Weaker players tend to emulate the best known successful players of the day, without having the skill set to support the style. Goren addressed this by developing a more conservative system for weaker players, which was popular for many years. However don't be misled to think that Goren played passively in tournaments; Goren played his own system very aggressively, because of possessing the intuition and skill to do so successfully.
Yes, there are in many hands low spot cards that one does not need to remember exactly; thinking of an x is sufficient. However, the bridge literature is chock full of examples where the careless play of a deuce or trey instead of 5 or 6 has blocked a key suit. One never knows for sure which low spots are irrelevant until the full hand is counted.
I remember well an argument between two mentors, during my formative bridge years, as to where two small diamond spots had been, on a hand played 4 years earlier. Yes, that was the key hand in an important hand in a lost final; but one cannot even attempt to remember that far back if one doesn't remember at the conclusion of the hand. (Corollary: they both agreed on the location of the other 50 cards.)
Remember this: As declarer, the opponents are using those spot values to communicate with each other; as defender, your partner is using those spot values to communicate with you, and you are using those spots to communicate with partner. Nothing is more frustrating than to lead a small spot to partner asking for the suit to be returned, and having partner return another suit instead allowing a FAMOUS declarer to make an unmakeable contract (Sorry Dan! Too good an anecdote to leave out; I'll buy you a beer.)