Want accountability? Then first you have to have countability.

As Police Scotland ends its second year, watchdogs - in Holyrood and beyond - are eager to show their teeth.

But to do so, to really hold the force to account, these watchdogs need hard numbers.

And, with Scottish policing coming together from nine different law enforcement agencies with nine different IT systems, such data is far from easy to find.

Take the row over stop-and-searches. The sheer volume of these interactions with the public - usually understood as frisking - has soared in recent years - even if it is currently subsiding dramatically.

That must be straightforward to count? No. Chief Constable Sir Stephen House has admitted the numbers could be out by "several hundred thousand" a year. That is a big deal.

To give a sense of scale, there were around half a million recorded in the first nine months of Police Scotland, during the period from January-December 2013, close to the peak of recording.

Understandably, real concern was raised over what impact these searches, two-thirds of which were officially deemed as voluntary or "consensual", were having on police relations with the public, especially teenage boys mostly likely to be stopped.

However, police insiders have always doubted the figures.

Glasgow's traditional "booze blitzes" - when drink is taken off kids in the streets - were counted as stop-searches, even though bottles and cans confiscated were in plain sight.

The Scottish Police Federation has long argued that the figures were somehow inflated by the drive to meet performance indicators.

The force says it has no targets for number of searches - but it does for what proportion of searches are "positive". This, say its critics, resulted in police declaring drink they found as a positive search. Police HQ disputes this.

But rank-and-file officers are not impressed. "The numbers are guff," declared an exasperated spokesman for the SPF after hearing Sir Stephen's admissions.

One of the big worries over stop-and-search was that youngsters under 12 were being frisked "consensually" despite widespread acknowledgement that were too young to be able to give consent.

This police agrees with this. This summer, said such stop and searches would stop. Cue more bad numbers.

The BBC used Freedom of Information laws to find out how many children of 11 or under were searched using a consensual search. The force refused to say, arguing its data were corrupted. But it was obliged to reveal its iffy figures, 365 in total.

The BBC published them in, admits the force, good faith. The result? A political row. On Friday, the force reported how many under-12s had in fact been frisked potentially in breach of its own policy. It was 18. That fact is now being independently checked.

When challenged at his main watchdog, the Scottish Police Authority (SPA), Sir Stephen, among other things, blamed the "clunkiness" of his IT systems, which are currently being revamped.

The police hadn't felt the need to tell the SPA about the BBC information request - or emerging numbers, even wrong numbers, on stop and search of children. Members of the authority grumbled; they should have been kept informed, they said. Sorry, said Sir Stephen.

Now the force has said Scotland should consider new police powers to search children for alcohol, making up for a "gap" left by their not being able to carry out consensual ones.

Sound reasonable? Maybe. But with numbers - and even the very definition of a stop-search, in such a mess, how can anyone be sure that such a move would be evidence-based?