How do tech teams work exactly?

[Gort11]

How much of a bonus does a tech team get for being specialised in a type of research?

For example:

Tech team 1 has skill 4 but is specialised in all of the things required for this particular tech.
Tech team 2 has skill 8 but has no specialisation.

Which is better?

While we're at it, how big a bonus to rocket techs and nuclear techs do rocket test sites and reactors give?

Is there a bonus for researching really old techs, like 1918 infantry in 1936?

 

[D3ATH]

I mostly get 3.4% or such per day on very old techs.

 

[Jamesk2]

There is a guide about techteam somewhere (perhaps in the HoI Wiki) so you can find it.
But by the way it's a rule of thumb to give the tech to the lv 4 techteam, as it costs you much less in money and you still have the lv 8 techteam for another job...

 

[Pepsi_max]

How much of a bonus does a tech team get for being specialised in a type of research?

For example:

Tech team 1 has skill 4 but is specialised in all of the things required for this particular tech.
Tech team 2 has skill 8 but has no specialisation.

Which is better?

While we're at it, how big a bonus to rocket techs and nuclear techs do rocket test sites and reactors give?

Is there a bonus for researching really old techs, like 1918 infantry in 1936?

The better is tech team 1 with all the things required.
Because this is faster (25-30%, something like that)
Cheaper (-50% cost)
You can put the lvl 8 techteam on another work

 

[Gort11]

It's more expensive to run tech teams that don't specialise? Mind blown.

 

[Jamesk2]

Lol an 8-skill techteam cost double when research than a 4-skill. It doesn't rely on specialization

 

[Limith]

Lol an 8-skill techteam cost double when research than a 4-skill. It doesn't rely on specialization

Skill really only matters for tech team cost and specializations. I generally would take a lower skill tech team if they have more specializations, unless the tech is way behind era, I have blueprints, or it is rocket tech (with me maxed out on test center), which I would pick a terrible team instead to lower money cost.

Look at the makeup of the tech, whichever component requires a higher number is also a factor.

 

[Rotten Venetic]

For speed, IIRC it was that 1 extra skill pt was worth two specialization.

Ex. A team with 3 specs and level 6 will be slower than one with 2 specs and level 7, but faster than one with 4 specs and level 5.

This is, of course, assuming all the components of the tech have equal requirements (or those that aren't are strict doubles, which can be easily counted twice). Also assuming that nuke reactor and rocket test site don't come into play. Generally speaking rocketry is kinda useless if you have the IC to build a good rocket test site (nuke techs not so much, because there's a lot of research that needs to be done with no or low reactor).

 

[JRHINDO]

This makes little sense... So a lousy team with lots of specs is better than a super team with no specs?
Equal research strength would be at minimum logical but....

 

[Mars6]

Having a specialisation in a certain tech gives you double research, so a tech 4 team with all specs equals a lvl 8 team, except for the cost In most cases, it is better to go for a lower tech team with more specs (I personally always take the teams that have more specs)

 

[am300307]

Meh there needs to be some sort of research assistant.

 

[Limith]

Meh there needs to be some sort of research assistant.

The old Hoi2/Arma editor had a research calc.

 

[Pepsi_max]

For speed, IIRC it was that 1 extra skill pt was worth two specialization.

Ex. A team with 3 specs and level 6 will be slower than one with 2 specs and level 7, but faster than one with 4 specs and level 5.

This is, of course, assuming all the components of the tech have equal requirements (or those that aren't are strict doubles, which can be easily counted twice). Also assuming that nuke reactor and rocket test site don't come into play. Generally speaking rocketry is kinda useless if you have the IC to build a good rocket test site (nuke techs not so much, because there's a lot of research that needs to be done with no or low reactor).

A level 5 with 4 spec will be faster than a level 7 with 2 specs.

 

[Rotten Venetic]

Man I tried so hard not to mix that up. Need more sugar.

 

[Pepsi_max]

Man I tried so hard not to mix that up. Need more sugar.

Well i could be wrong, i'm saying that from my memory. But if i'm wrong that's not by far. And the cost of the level 7 tteam will be quite superior to the level 5 tteam. Rendering it non-profitable.

 

[Rotten Venetic]

No, Pepsi, you're right. At 2 specialties vesrus 4, the team would have to have 9 skill points vs 5.

I just did a little test, with Poland:
1 Jan 1936:
Gave Centralny Okreg Przemyslowy the task of researching Accounting Machine (1928), first component is Mechanics, level 8. The team is skill 4 and has this spec.
Gave Stocznia Gdyna (no1 ever uses them) the task of researching Improved Construction Engineering (1930), first component Industrial Engineering level 8. The team is skill 4, without the spec.
Both these techs are well within the maximum behind-the-times research bonus, which is 3 in my Misc.txt.

As soon as January the 5th it was obvious that COP was keeping steadily at 2 times the speed of Stocznia Gdynia (6% vs 3%), and in my misc.txt file, the blueprint bonus is at 3, so we can conclude that it does not affect the specialty bonus, which remains at 2. There were no blueprints involved.

Second test, again POL 1 january 1936:

Used the same techs - Accounting Machine and Improved Construction Engineering.

COP gets Imp. Con. Engineering again (reminder: skill 4, has specialty, difficulty 8)

This time, CWS researches Accounting Machine. They are skill 3 and have the Mechanics specialty for the first component.

By January 13, COP has 18.2% of the tech done, and CWS, 16.4%, giving a difference of 10.(97560)% in favor of COP.

Further tests involving Ethiopia and the Soviet Union reveal that, for difficult techs and poor teams, the difference in skill points is more pronounced (up to 13.6% for diff. 8 tech components between level 1 and 2 teams with no specs on the first component), or less (down to 10.86% for level 6 and 7 specs, and difficulty 2 tech components).

This means that, in fact, if team A is one specialty behind but 2 skill points ahead of team B for a given tech, A will actually research it faster by about 4-6% depending on the tech



TL, DR: An extra specialty gives around +20% and an extra skill point gives around +11%, depending on the difficulty of the relevant component/s.

More difficult relevant components relative to the tech's other components favour specializations.
More difficult relevant components relative to the teams' skill points favour higher skill points on the tech team.

 

[Pepsi_max]

Also, it seems to be slighly better to choose components which are researched the last (even if all the costs are same). Something like 1-3% difference.

 

[Gytis]

The simple calculation I due when deciding between teams is multiplying the skill level with each of the research component difficulty, and doubling the component points if its a teams specialty.

Heres an example: for a 6 5 3 5 10 technology there are two possible teams, 4 and 5 skill respectively, but the first one being specialist in the "10" field. So the the first one, it would be 4*6+4*5+4*3+4*5+(4*10)*2 equal 156 points, and the second one should get 5*6+5*5+5*3+5*5+5*10 equal 145 points. In this example, a 4 skill team with a specialty would research a bit faster than a 5 skill team without specialties.

 

[Rotten Venetic]

Yes, pretty much like that, I guess. The tech difficulties' weight is an important consideration, though for the tech in question it's still very close to the average of 33% for the final one.

 

[demanvanwezel]

Also, it seems to be slighly better to choose components which are researched the last (even if all the costs are same). Something like 1-3% difference.

probably because progress goes bar per bar, if you get 3% a day and you only need 2% more to fill the component bar then the last 1% is lost