TRACOM: Is it worth it?

[Ian R]

ORIGINAL: RangerJoe

But the Japanese national pilot experience average goes down later in the game, does it not? So by getting them early, you are getting them at the higher national average. Train them to be good Kamikaze pilots instead of poorly trained ones.

I have a recollection that someone, possibly Alfred or the Elf, stated that just because you graduate some pilots from 'training command' early, does not mean you get more raw no-name recruits into the system to replace them. So at best you get them onto the map earlier so as to put them in an "OTU" and polish them up some more. Or throw them in the deep end.

[RangerJoe]

I understand that you don't get more, but if you pull them out before the national average at graduation drops then they should have a higher experience.

[jdsrae]

There is a supply vs HI point trade off consideration, but...
HI is paid based on 5 x number of pilots, I believe once per month.
Supply is burnt by the number of planes in the air units used for training.
For the IJN, USN, RN, supersized air units can have 100 pilots sharing 2 planes.
Using air units with 2 planes uses a lot less supply than units with full plane quotas. (At the risk of sparking the number of planes debate again!)

I don’t have any pilots in training command yet, but there may come a time when I move some IJN fighter pilots there.
It looks like it will be a long time before I am able to drain the IJA general reserve pool, so I can’t see the point in rushing IJA pilots through their 12 month basic course just to leave them in the inactive reserve pool.

[Alfred]

ORIGINAL: Ian R

ORIGINAL: RangerJoe

But the Japanese national pilot experience average goes down later in the game, does it not? So by getting them early, you are getting them at the higher national average. Train them to be good Kamikaze pilots instead of poorly trained ones.

I have a recollection that someone, possibly Alfred or the Elf, stated that just because you graduate some pilots from 'training command' early, does not mean you get more raw no-name recruits into the system to replace them. So at best you get them onto the map earlier so as to put them in an "OTU" and polish them up some more. Or throw them in the deep end.

Correct.

The monthly intake to flight training school is fixed for all nationalities. It does not increase to cover a school deficit resulting from excess withdrawals.

Alfred

[BillBrown]

From the Pilot Management Addendum:

Groups can fly normal Training missions (with a training percent) which occur in the
AM and PM air phases. These gain both skill and experience points.
Groups will also gain skill and experience (after passing a training check against the
training percent of the group) at the end of each day:
1. if the pilot’s experience is less 50 (plus pilot’s missions and kills) and less
than the overall group experience level
2. if the pilot’s experience is less 50 (plus pilot’s missions and kills) and less
than the leader’s skill
3. if pilot is in a dedicated Training group with some Instructors (adds higher
increments to accumulators)
4. if pilot is in a dedicated Training group without Instructors (higher success
and slightly higher increments than a normal group)
5. if pilot is in a group with some training percent (number of veteran pilots
(experience of 80+) increases chance of successful training)

If a scenario designer makes some dedicated training groups, they will train pilots
faster than normal air units that are given training orders.

[Alfred]

The private dev forum discussion on dedicated Training groups makes for very interesting reading.  There was a very sound reason why post release the official decision was to not have any in the official scenarios.
 
Private scenario designers of course can do how they please but I would not recommend incorporating them.
 
Alfred

[Kull]

ORIGINAL: RangerJoe

But the Japanese national pilot experience average goes down later in the game, does it not? So by getting them early, you are getting them at the higher national average. Train them to be good Kamikaze pilots instead of poorly trained ones.

No it does not. Stays constant from 1941 thru 1946, for both IJA and IJN. The table is in Post #10 of this thread.

[RangerJoe]

ORIGINAL: Kull

ORIGINAL: RangerJoe

But the Japanese national pilot experience average goes down later in the game, does it not? So by getting them early, you are getting them at the higher national average. Train them to be good Kamikaze pilots instead of poorly trained ones.

No it does not. Stays constant from 1941 thru 1946, for both IJA and IJN. The table is in Post #10 of this thread.

My mistake, I thought that it went down. It did historically if I remember correctly.

[Kull]

ORIGINAL: awaw

I believe what rustysi is saying is that:
1. Normally a pilot has an average 30 air skill.
2. By forcibly “graduating” him before that, he will have ,say, 20 air skill.
3. Somehow, somewhere, you will have to train him on-board, and taking him the extra distance from 20 to 30 skills costs supplies.
4. 20-30 skill is fast, but still a supply cost somehow

Before responding, it's important that we look at the in-game training status visuals and (in combination with a few quotes from the coder & Alfred) try to determine what they mean. The attachment is the the Scenario 1 Japanese pilot training graphic as of 12/9/41. It shows pilot classes grouped by quarter, in which the "Average Experience level" is that of all pilots in three different classes. Each month, a new class is inducted and one of the classes in each quartile will carry it's unknown average into the next quartile, while the 12th month class (and it's unknown average) will graduate.

Accordingly, when the attachment tells us that the last IJA class has an average of 29, you can bank on the fact that "class 12" is higher, but that's all you can know. The Devs probably used this mechanism to obscure the specifics of the data for each class, so we can never know the average number of any one specific class (much less the distribution of per-pilot experience levels within each class).

The details are missing, but there are some things we can infer. Let's start by looking at the average experience for IJA pilots in each quartile:
1st Quartile group: 1 (+1)
2nd Quartile group: 12 (+11)
3rd Quartile group: 22 (+10)
4th Quartile group: 29 (+7)

Don't focus too much on the specific averages or even the rates of increase - it varies a LOT from one game and one time period to another. But what they all have in common (absent a ransacking of pre-graduate pilots by a desperate player), is that the total increase will be at least high 20's by the 4th Quartile. And remember that for many of the Allied nationalities (especially in the late war), the numbers will increase from single digits into the 40's. Thanks for bearing with me, because here's where that leads:

Quote #1:

ORIGINAL: michaelm75au

Adjustments to the pilot training happens on the first day of the month.
The monthly EXP increase of the raw pilots (ignoring TRACOM effect) is a random amount (-2 to +2).

If you just do the math - +2 as the max EXP increase per pilot per month over 12 months - there's no way you can get from 1 to 29 in a year. The luckiest pilot, who randomly gains +2 every single month, would wind up at 24. And what about all the guys getting a -2? This HAS to be a mechanism that derives a spread between the experience levels of individual pilots in each class, not the basis by which the entire class goes from 0 to "something less than 35" in 12 months. That mechanism is probably this:

Quote #2:

ORIGINAL: michaelm75au

The new monthly EXP starts out as an average of the pilots in the month and those moving from the previous month.

Although I can't wrap my head around the math which would make it work, one way or another the algorithm drives overall pilot experience higher than the +2/-2 mechanism (especially in classes 4-12). So how does TRACOM fit in? Two ways - the first of which relates directly to Quote #2 above:

Quote #3:

ORIGINAL: michaelm75au

The number of TRACOM guys can increase the month's EXP rating of the raw pilots.
If the new EXP exceeds the national EXP, then a random number of pilots will 'graduate' early.

A rough guide is: for every 10 in TRACOM (of the correct nationality), the monthly EXP might be increased by one.

Read that carefully and then read it again. He's not talking about quartiles or the graduating class per se, but rather "the month's EXP rating of the raw pilots". Whether that's "per pilot" or "per class" is not clear, but either way it happens to each class-or-pilot EVERY MONTH. The operative word is "can", so it's not guaranteed, but 12 times is a LOT of opportunities for increase. In particular it's huge if this drives the overall class number higher in a month, because (according to Quote #2) it affects not just this class but all those "moving from the previous month".

For the second TRACOM benefit, we have this :

Quote #4:

ORIGINAL: Alfred

Graduating pilots come out with experience which is within a range +/- of the national average. Having pilots in Tracom will aid reducing the negative spread but will not boost them to above the national average.

Alfred

This statement clearly applies to the +2/-2 effect (see Quote #1) which is intended to expand the experience delta between the pilots in EACH class, not just the graduating class. In particular it "reduces the negative spread", which - at the most simplistic level - would have to mean that it's driving the overall average experience higher. Is that per month, per pilot, per class? Don't know, but it's not a bad thing, and almost certainly contributes to the Quote #2 mechanism.

Which finally brings us back to this:

1. Normally a pilot has an average 30 air skill.

- We don't actually know what that number is (as explained above), but for sake of argument, OK.

2. By forcibly “graduating” him before that, he will have ,say, 20 air skill.

- Absolutely not. As shown above, the TRACOM mechanism doesn't work by slashing months off the graduation time of some random individual or even an entire class (which seems to be your assumption), but rather looks to see "if the new EXP exceeds the national EXP, then a random number of pilots will 'graduate' early". The guys graduating early CANNOT be "20" because they are part of a class which has an overall average that "exceeds the national EXP" (not something that's going to happen if the class is populated by pilots with a "20" EXP level)

3. Somehow, somewhere, you will have to train him on-board, and taking him the extra distance from 20 to 30 skills costs supplies.
4. 20-30 skill is fast, but still a supply cost somehow

- Both statements are false, since they follow-on from the assumption in #2 that the average experience of "early graduates" is less than that of those who spend 12 months in training. In fact, as per Quote #4, the reverse is true since the 2nd TRACOM mechanism improves the experience numbers for the guys at the bottom of the graduating class. Which means that graduates assisted by TRACOM should require LESS, not more, supply to train further.

[awaw]

@Kull

Thank you for the elaborate response, but I think we are on 2 different channels here.

when I say "forcibly graduate early", I was referring to this sequence:
a. A pilot normally stay in the training program for 12 mths.
b. If a player forcibly "use up" pilots that graduated from the training schools, and start draining pilots from (say) 10-12 mth cohort, they are denying these pilots "free" training that could have accumulated during the remaining period. While not all pilots will benefit from staying in flight school for the full period, statistically, some will gain in xp. Drawing from the 10-12 mth cohort is what I mean by "forcibly graduate early"
c. Training pilots on map cost supplies, and putting a 10-12 mth pilot cadre instead of a normally graduated pilots for on-map training will imply a supply cost.
d. a~c does not care about how the game moves pilots between the (1~3),(4-6),(7~9) and (10~12) groups. Only point of consideration is that statistically, a pilot cadre in the (1~3) group will have less skill points than the (4~6) group, and the time spent on-board to improve the pilot from (1~3) to (4~6) costs supplies.
e. So there is still a HI to supply tradeoff.

[HansBolter]

All this conjecture makes my head hurt.

Having to worry over HI/supply trade offs is the best argument for NOT playing the Japanese side.

[Ian R]

ORIGINAL: HansBolter

All this conjecture makes my head hurt.

Having to worry over HI/supply trade offs is the best argument for NOT playing the Japanese side.

It certainly is a simpler proposition for the allies - get as many replacement pilots into as many restricted groups as possible, and train, train, train.

[Kull]

ORIGINAL: awaw

@Kull

Thank you for the elaborate response, but I think we are on 2 different channels here.

when I say "forcibly graduate early", I was referring to this sequence:
a. A pilot normally stay in the training program for 12 mths.
b. If a player forcibly "use up" pilots that graduated from the training schools, and start draining pilots from (say) 10-12 mth cohort, they are denying these pilots "free" training that could have accumulated during the remaining period. While not all pilots will benefit from staying in flight school for the full period, statistically, some will gain in xp. Drawing from the 10-12 mth cohort is what I mean by "forcibly graduate early"
c. Training pilots on map cost supplies, and putting a 10-12 mth pilot cadre instead of a normally graduated pilots for on-map training will imply a supply cost.
d. a~c does not care about how the game moves pilots between the (1~3),(4-6),(7~9) and (10~12) groups. Only point of consideration is that statistically, a pilot cadre in the (1~3) group will have less skill points than the (4~6) group, and the time spent on-board to improve the pilot from (1~3) to (4~6) costs supplies.
e. So there is still a HI to supply tradeoff.

Let's review:

- Fewer pilots in the 12 training classes means less expenditure of HI. That we seem to agree on.
- The pilots who graduate early, have an experience level which is close to or at the national experience level. Keeping them in the training class does not get them any additional experience. It simply burns HI.
- Graduates, early or otherwise, go into the "pool of pilots without a name". They do not appear in units and cannot be trained further in ANY WAY until the player draws them into a unit. Thus adding more pilots to the "un-named pilot pool" has NO EFFECT on supply.
- Lastly, airframes burn supply, not pilots (see manual, pg.252). So even if you did drag all the new pilots straight into a unit, it won't have any effect on supply.

In sum, TRACOM is NOT a trade-off between HI and supply. Rather it reduces the expenditure of both.

[RangerJoe]

ORIGINAL: Kull

ORIGINAL: awaw

@Kull

Thank you for the elaborate response, but I think we are on 2 different channels here.

when I say "forcibly graduate early", I was referring to this sequence:
a. A pilot normally stay in the training program for 12 mths.
b. If a player forcibly "use up" pilots that graduated from the training schools, and start draining pilots from (say) 10-12 mth cohort, they are denying these pilots "free" training that could have accumulated during the remaining period. While not all pilots will benefit from staying in flight school for the full period, statistically, some will gain in xp. Drawing from the 10-12 mth cohort is what I mean by "forcibly graduate early"
c. Training pilots on map cost supplies, and putting a 10-12 mth pilot cadre instead of a normally graduated pilots for on-map training will imply a supply cost.
d. a~c does not care about how the game moves pilots between the (1~3),(4-6),(7~9) and (10~12) groups. Only point of consideration is that statistically, a pilot cadre in the (1~3) group will have less skill points than the (4~6) group, and the time spent on-board to improve the pilot from (1~3) to (4~6) costs supplies.
e. So there is still a HI to supply tradeoff.

Let's review:

- Fewer pilots in the 12 training classes means less expenditure of HI. That we seem to agree on.
- The pilots who graduate early, have an experience level which is close to or at the national experience level. Keeping them in the training class does not get them any additional experience. It simply burns HI.
- Graduates, early or otherwise, go into the "pool of pilots without a name". They do not appear in units and cannot be trained further in ANY WAY until the player draws them into a unit. Thus adding more pilots to the "un-named pilot pool" has NO EFFECT on supply.
- Lastly, airframes burn supply, not pilots (see manual, pg.252). So even if you did drag all the new pilots straight into a unit, it won't have any effect on supply.

In sum, TRACOM is NOT a trade-off between HI and supply. Rather it reduces the expenditure of both.

Excellent summary.

[geofflambert]

All this math makes my head hurt, so I don't bother worrying about it. What engages me most is how to get the proper balance of skills in the right number of pilots for the right number of aircraft of each type. For instance, excluding fighter pilots (and FPFs are another complication, as they can both search and do ASW) for the IJN I consider naval search to be a nominal skill, one of the first things everyone is trained in. Then comes naval bombing, but occasions have arisen where I experienced a shortage of training in naval torpedo attacks and had to pad front line units with crews that are well trained in that but may not be in delivering bombs on ships or in naval search. I was able to at least do that because I began training some bomber recruits in torpedo attack before anything else. As a result I would tend to restrict ranges to standard range, assure that sufficient torp inventory was available and that sufficient search aircraft were working in conjunction with torpedo bombers. That raises a question I've wondered about, if 20% of your crews are adept at search and you order 10% to search, will only experienced crews get that assignment or is a "random" generator used? The same question would apply to other combinations of skills. I depend more on TB crews to do search and tend to depend on DB crews more for ASW, but eventually both DBs and TBs do search as well as ASW. For the IJN I generally get ASW training done aboard carriers in port or otherwise idle, and crews which are constantly searching inevitably gain ASW skill, albeit slowly (That goes for naval bombing as well, as I tend to have most Emily and Mavis squadrons constantly searching and without benefit of torp inventories, and in some cases Nell squadrons as well. Since the IJA problems are much simpler and the supply of crews more numerous I try and have them take as much of the burden as possible in search (Dinahs mainly) and ASW (Lillys mostly) and hold back IJN squadrons, both F and B, out of bloodbaths if I can, and maximize the opportunities to be aggressive with IJA squadrons. Etc. etc. etc.

[mind_messing]

ORIGINAL: Alfred

The private dev forum discussion on dedicated Training groups makes for very interesting reading.  There was a very sound reason why post release the official decision was to not have any in the official scenarios.

Private scenario designers of course can do how they please but I would not recommend incorporating them.

Alfred

Reading between the lines: too easy to stack groups full of high EXP pilots leading to very short training cycles?

Follow-on question: you mentioned that they are in no official scenarios. Is it the case that the late war groups that arrive with the tag of "Training group" and are restricted to training missions do not follow the rules above?

[Zorch]

ORIGINAL: geofflambert

All this math makes my head hurt, so I don't bother worrying about it...

Excellent philosophy of life!

[RangerJoe]

ORIGINAL: Zorch

ORIGINAL: geofflambert

All this math makes my head hurt, so I don't bother worrying about it...

Excellent philosophy of life!

An accountant was being interviewed by an executive at a company for a job there. The executive asked "What is two plus two?"

The accountant did not respond immediately but got up, closed the windows, and shut the blinds. Then he went to the door, shut it, and locked the door. Then he sat back down in his chair and whispered "What do you want it to be?" [:D]

But I can prove on paper that half of one is one, half of two is two, half of three is three, half of nine is four, half of ten is five, half of eleven is six, half of twelve is seven, and half of thirteen is eight.

[Ian R]

ORIGINAL: RangerJoe

and half of thirteen is eight.

In base 8, base10 13 = 15

half of base8 15 = 7.4, but that rounds up to one bigger bead on a base8 abacus.

So in base 8, that is a 10, which in base 10...

is 8. [;)]

QED

Edit - 7.4 is visually confusing - think of it as 7 and 4/8ths.

[RangerJoe]

ORIGINAL: Ian R

ORIGINAL: RangerJoe

and half of thirteen is eight.

In base 8, base10 13 = 15

half of base8 15 = 7.4, but that rounds up to one bigger bead on a base8 abacus.

So in base 8, that is a 10, which in base 10...

is 8. [;)]

QED

Edit - 7.4 is visually confusing - think of it as 7 and 4/8ths.

Nope, nothing that complex. You are thinking too hard. [;)]