This issue is not a blocker for the release of V1.1.0, but we should address it afterward.

Change 565459 had a related patch set uploaded (by Gergő Tisza; owner: Gergő Tisza):
[mediawiki/extensions/GrowthExperiments@master] Use queue length, not total result count for initation dialog result counter

https://gerrit.wikimedia.org/r/565459

Can't reproduce the weird part - selecting Zdraví + Společenské vědy gives me 4 hits. Other than that, this is just due to (not) double-counting as you say. Which is currently somewhat incorrect as the tasks do get duplicated (the patch makes sure that we always report the number of task cards, not the number of matching articles, when the two differ), but once T243036: Newcomer tasks: rules for duplicate results is fixed this will be the correct behavior.

Change 565459 merged by jenkins-bot:
[mediawiki/extensions/GrowthExperiments@master] Use queue length, not total result count for initation dialog result counter

https://gerrit.wikimedia.org/r/565459

Re-tested https://gerrit.wikimedia.org/r/565459 - the test cases from task description show the same incorrect results. I corrected the case for - it should 4, not 7 as it was reported before: (correction) Zdraví + Společenské = 4

This task is a sub task for T243036: Newcomer tasks: rules for duplicate results; moving it back to Ready for Development to wait on T243036.

This seems to me like a duplicate of T243036 but maybe I am misunderstanding the issue. Is there anything here that would be left after T243036 is done, of that could be done without resolving T243036?

In T242814#5821500, @Tgr wrote:

This seems to me like a duplicate of T243036 but maybe I am misunderstanding the issue. Is there anything here that would be left after T243036 is done, of that could be done without resolving T243036?

Pinging @MMiller_WMF
I re-tested this issue - and it seems that actual count (the count of unique articles) is correct (see the details of testing below).

However, before closing this task as a non-issue, let's consider what users' expectations should be when a user browses through topics selection and looks at the counter? If I combine topics why I should see less number of articles that the sum of numbers of articles for each of the topics? Maybe to improve users' experience the counter should say "N unique topics"?

https://cs.wikipedia.beta.wmflabs.org/wiki/Speci%C3%A1ln%C3%AD:API_p%C3%ADskovi%C5%A1t%C4%9B#action=query&format=xml&list=growthtasks&gttasktypes=links&gttopics=business%7Ceconomics

Count for unique articles=10

<growthtasks totalCount="14">
      <suggestions>
        <suggestion title="Ekonomický sektor" tasktype="links" difficulty="easy" order="0">
          <topics>
            <_v>
              <_v>economics</_v>
              <_v>100</_v>
            </_v>
          </topics>
        </suggestion>
        <suggestion title="Ekonomika Robinsona Crusoe" tasktype="links" difficulty="easy" order="1">
          <topics>
            <_v>
              <_v>economics</_v>
              <_v>50</_v>
            </_v>
          </topics>
        </suggestion>
        <suggestion title="Mexická ekonomická krize" tasktype="links" difficulty="easy" order="2">
          <topics>
            <_v>
              <_v>economics</_v>
              <_v>12.5</_v>
            </_v>
          </topics>
        </suggestion>
        <suggestion title="Helikoptérové peníze" tasktype="links" difficulty="easy" order="3">
          <topics>
            <_v>
              <_v>economics</_v>
              <_v>11.111111111111</_v>
            </_v>
          </topics>
        </suggestion>
        <suggestion title="Milton Friedman" tasktype="links" difficulty="easy" order="4">
          <topics>
            <_v>
              <_v>economics</_v>
              <_v>16.666666666667</_v>
            </_v>
          </topics>
        </suggestion>
        <suggestion title="Průmyslová politika Evropské unie" tasktype="links" difficulty="easy" order="5">
          <topics>
            <_v>
              <_v>economics</_v>
              <_v>10</_v>
            </_v>
          </topics>
        </suggestion>
        <suggestion title="Ekonomika měst a obchod ve středověké Anglii" tasktype="links" difficulty="easy" order="6">
          <topics>
            <_v>
              <_v>economics</_v>
              <_v>20</_v>
            </_v>
          </topics>
        </suggestion>
        <suggestion title="Kapitalismus dohledu" tasktype="links" difficulty="easy" order="7">
          <topics>
            <_v>
              <_v>economics</_v>
              <_v>25</_v>
            </_v>
          </topics>
        </suggestion>
        <suggestion title="Ekonomika Svazijska" tasktype="links" difficulty="easy" order="8">
          <topics>
            <_v>
              <_v>economics</_v>
              <_v>33.333333333333</_v>
            </_v>
          </topics>
        </suggestion>
        <suggestion title="Ekologická daň" tasktype="links" difficulty="easy" order="9">
          <topics>
            <_v>
              <_v>economics</_v>
              <_v>14.285714285714</_v>
            </_v>
          </topics>
        </suggestion>
      </suggestions>
    </growthtasks>

Honestly, I don't think anyone will care about it. Search result counts are often a bit strange since counting large sets are hard and search engines tend to cheat in various ways (try paging search results in Google or Gmail for example and watch how the counts fluctuate).

In T242814#5830760, @Tgr wrote:

Honestly, I don't think anyone will care about it. Search result counts are often a bit strange since counting large sets are hard and search engines tend to cheat in various ways (try paging search results in Google or Gmail for example and watch how the counts fluctuate).

I agree (after thinking more about it). It'd be interesting to hear some users' feedback if there would be any.