Another possibility could be that simply your experimentation had flaws.
This is not a direct criticism pointed at your own methods, but rather an explanation on why it might be problematic to compare in-game climb performance with real-life performance by engaging "auto-climb".
For one thing, auto-climb is not known to be the most efficient method of climbing in Aces High. Despite popular beliefs "auto-climb" is actually "auto-speed" - the plane will stabilize its general heading and its balance in the roll axis, while adjusting its pitch to keep the plane stable at a certain set speed.. and therefore, it does not take into account a number of factors that might effect climbing efficiency. The optimal speeds for climbing might change throughout the range of altitudes a plane travels - which, is clearly more than the "auto climb" feature might handle.
Therefore, my guess is that the auto-climb can provide an objective reference of climbing capabilities only when it is directly compared with other in-game planes also using auto-climb.
ie) When one compares the Bf109 with the F4U, the difference in magnitutde of climbing capabilities between the two planes will be presented at a relative scale: one will first engage the Bf109 in auto-climb, and then compare it with the F4U engaging in auto-climb. It is highly probable that both the 109 and the F4U will not match real-life figures per se, but rather only the relative difference between the two planes engaged in auto-climb, will match that of the relative differences between real-life figures.... for instance, if real life climb figures show a 30% advantage in climb in favor of the 109, the same 30% ratio will show between the auto-climb results compared, despite the ultimate "time to reach X altitude" figure might not match that of real-life.
The only way to actually try and match real-life climbing figures with that of the game, IMO, would be to look up historical documents and see what sort of methods the actual planes used, to perform climb tests. One particularly necessity would be try to find out the average climb speed the plane used in actual real-life testing, and adjusting AH "auto-climb" speed to match that.
If there's one thing I've found out during my own series of comparative tests in turn performance, it is that coming up with an objective and mechanical testing method is incredibly difficult. My suggestion is that you should try to consult these boards and have a discussion in which testing method might provide the results you are seeking... before any testing is actually done in the first place.
To the extent of my knowledge, every piece of information we have for AH planes were actually tested by different people with different test methods - which seriously questions the objectivity of it all.
As a side note, I myself do not trust other people's results in turn performance testings, for example, because I find serious faults in the objectivity of their testing methods - whereas my own tests were devised to set each of the plane's to their mechanical limits and leave out as much human factor as possible, others were rarely so thorough.
A stark example of why a totally objective method of testing is required can be observed from one of my own test notes, linked in my sig below; "The AH Compendium of...."