Alternate Tech Engine Suggestion

[John DiFool]

I have been carefully mulling over some alternatives

to the current tech system. Most people seem to

agree that it is too dependent on luck. Witness

the recent thread on "Endgame AI Weirdness" for

an example-in my last game the Russians got up to

level 5 jets AND level 5 infantry within two years,

while I, despite having no less than 8 Tech Points,

got just TWO advancements over the same period.

This can lead to some wild imbalances (esp. in

PBEM play).

My suggestion then is to reduce (but not totally

eliminate) the role of luck in getting tech.

How this would work is via what I call Tech

Increments (for lack of a better term). Basically

for each level of tech invested, you will get a

certain number of Increment Points per turn. When

these IP total up to 100, you get an advance.

Here is the table, then I'll provide an example:

Tech Points....Base Tech %....Tech Increments/Turn

1________________5%_____________2-8

2_______________10%_____________5-15

3_______________15%_____________8-22

4_______________20%_____________11-29

5_______________25%_____________15-35

[the message editor doesn't recognize extra spaces- sorry ]

Say you invested 4 points in Jets (all at once).

On the first turn you will get between 11 and 29

Increment Points, and so on for each succeeding

turn (unless you change it), until the total

reaches 100, at which point you finally get the

advance.

Odds are that it will take you 5 turns-but it

could take 4. In any event you are guaranteed to

get it within 10 turns (and that is very unlikely-

requiring NINE straight 11's).

You could have 'extra' IPs carry over towards the

next advance. You could also make them 'secret'

(i.e. not visible to the player, who has to then

guess as to when he will get the advance, thus

limiting the utility of all those Perfect Plans :cool: ).

What this does is limit the role of luck in the

gaining of tech (while not eliminating it). No

longer will you get zilch over the space of 12

turns-but nor will you get 3 advances in 6 turns.

This also will make cheating either impossible or

VERY obvious. And the best thing is that it

doesn't complicate the game or add some new

wrinkle that the AI can't handle: in fact if

Hubert decides to hide the IPs from the player it

will APPEAR to work the same as before.

Thoughts? Hubert?

John DiFool

[Desert Dave]

As originally posted by John DiFool:

What this does is limit the role of luck in the

gaining of tech (while not eliminating it).

Interesting approach, but I like the idea of random tech advances since it makes each game unique. :cool:

Or, in order to make the effects less wild and haphazard, you could have a house rule with your opponent, as I have done.

The issue of cheating is not an issue for me (... I try not to worry over things I cannot control; if someone beats me 9 times in a row, then I will briefly wonder, but otherwise... ), but I can see where and why you would be concerned.

Tone it down a bit? Perhaps. But not to the extent that each game resembles every other.

[Hubert Cater]

This is a very good idea and an alternate approach, but I think I would most likely decide against it unless it was slightly modified. As Immer has alluded to, it may cause games to be very similar in the long run without randomness.

I guess this is what the current research tech boils down to, two major camps, one that prefers the randomness of the research tech and one that does not. My design philosophy for this game is to go against the guaranteed reward for investment and in the end this boils down to preference.

Now I am not saying that the random research tech is perfect, but I will most likely stay away from the suggestions that lean towards the set return on investment or the ones that will recreate tech advances (albiet with slight variance) exactly as they happened in the war (as others have suggested in the past).

Hope this helps,

Hubert

[John DiFool]

Originally posted by Hubert Cater:

This is a very good idea and an alternate approach, but I think I would most likely decide against it unless it was slightly modified. As Immer has alluded to, it may cause games to be very similar in the long run without randomness.

snip

True, there is always the fun factor to consider

here (I too look eagerly towards getting advances-

only to shake my fist at the monitor when I get

zilch :mad: ). I guess I am more concerned with

game balance-sometimes you can find a way to win

when the random device gives you the short end of

the stick, but mostly, if your opponent lucks out

in the tech arena, and you don't, you will likely

lose.

Plus games won't be the same if people

take a different tack in their investments from

game to game: both in terms of how much they

spend, and where. Unless there is a given Perfect

Plan to follow for each possible grand strategy

(i.e. as the Germans, pump Ind Tech up if you

plan to invade Russia, go for Jets if doing a Sea

Lion first, etc.).

John DiFool

[ September 02, 2002, 06:27 PM: Message edited by: John DiFool ]

[pzgndr]

The randomness doesn't concern me as much as the 10 point max and 5 per area limits which, if maxed out, lead to some unrealisticly rapid advances in either system. Some way to allow players to set these limits at game start, either on the options menu or an editable config file, needs to be considered so we can set lower limits and slow down the advances some. House rules work for two players, but don't help when playing the AI.

As for game-to-game research strategies, perhaps variants which give sides particular random advantages could be considered, like a virtual research point in an area. If Italy had a bonus for rockets one game, maybe they would pursue that line. Nothing significant, just a random boost.

[VictorH]

What about using the set return on investment as an "Option"? I.E. give players a choice, but make both sides agree to it.

[ September 02, 2002, 11:28 PM: Message edited by: VictorH ]

[Bruce70]

All research gets you closer to a breakthrough IRL. Even if the current direction of the research will ultimately lead to a dead end that is still one path that has been eliminated. If you are at all competent that path will not have to be tried again, hence you are closer to a breakthrough. The unknown is not knowing how long a breakthrough will take, ie how many paths you will have to try.

So a more realistic approach would be to have a research counter that is set to a random amount as each tech is achieved. Then the counter is reduced by the amount of investment until it reaches 0 at which time you get the next advance.

If the initial value is large you may never get an advance, if it is small you may get an advance very quickly. The player doesn't know what the initial amount is so they don't know how long it will take - but they DO know they are getting closer. Also if you stop research in an area and then start again you haven't lost the work you have already done.

[ September 03, 2002, 01:05 AM: Message edited by: Bruce70 ]

[Willard]

Hubert-

Presently, you have to spend 250 MPPs for a research point. You can than take that research point and invest in Industrial Tech. I think once you reach level 1 Ind Tech, that research point is "gone." Essentially keep the randomness of when the research advance occurs. However, dont allow for reinvestment of that point. Presently, there is no cost associated with the advance except the time the advance took to be "discovered." If one invests that 250 MPPs and gets the level 1 Ind Tech, then thats it. You should not be able to to get that point back to reinvest.

For example, you invest 5 research pts in Industrial Tech. You should get the following sliding scale affect:

MPPs Res Pts Tech Level Time

1250 5 0 Random

1000 4 1 Random

750 3 2 Random

500 2 3 Random

250 1 4 Random

0 0 5 Random

Obviously since you cannot "reinvest" tech pts, you may have to tweak (i.e. slightly lower to 225 or 200) the costs for such pts. Or you could make research pt cost decrease with each higher level of Ind. Tech!!!

However, I think it is necessary that you dont reuse research pts. Once the advance is made you should not be able to recoup the MPPs and reallocate elsewhere.

[ September 03, 2002, 02:51 AM: Message edited by: Willard ]

[BlueMax 1939]

I think Bruce70's idea was by far the best. Im all for it. Not a big fan of the system as it is because it is so easy to skew things out of proportion; i think it has a huge impact when you can get extremely lucky or unlucky...

[zappsweden]

Proposals for the research system.

1) Please include an OPTION for "semi-random research".

2) Change research cost to 200 but make research cost some maintenance (3MP per point).

3) Make a slight "catch up effect". It should mean that from level 2 onwards the tech level should not allow 5 research points.

Research ¤¤¤ Maximum research points allowed

level 0 ¤¤¤ 5

level 1 ¤¤¤ 5

level 2 ¤¤¤ 4

level 3 ¤¤¤ 3

level 4 ¤¤¤ 2

level 5 ¤¤¤ no points

MY SEMI-RANDOM SYSTEM:

The most important thing with a semi-random

system is that the MAXIMUM time for advance is somewhat practically restriced (by probability) and that rapid advances also are possible even whith small research funds. A research with 5 points should take 4 turns but probably it may vary from 1-8 turns. The progress should be hidden and be rated from 0-100%.

Also, a low point research should be able to advance quickly so that it is still unpredictable. The spreas must be high otherwise high research points will ALWAYS beat low research points. The random system used currently has too much spread which means that it affects the war outcome too much.

Every turn the progress increases. A base value is added and sometimes a bonus value is added (roughly 13% chanse). The bonus will be small (the first value) or large (the last value) decided on random . 65-35 odds in favor of the first value could be used.

research points ¤¤¤ base value ¤¤¤ bonus values

1 ¤¤¤ 0-7 ¤¤¤ 25 or 95

2 ¤¤¤ 0-16 ¤¤¤ 50 or 100

3 ¤¤¤ 0-25 ¤¤¤ 75 or 100

4 ¤¤¤ 0-32 ¤¤¤ 100 or 100

5 ¤¤¤ 0-40 ¤¤¤ 100 or 100

In the semi-random system a country using 3 research point will on average get a tech increase once every 7 turns but it practically will be in the range 1-12. That 3 research points take over 15 turns will then be almost impossible with the new system contrary to the old random system where it has slight possibilities.

[ September 04, 2002, 07:14 PM: Message edited by: zappsweden ]

[Rediroc]

Well something should be done to limit the reseach accomplishable in one year. A very lucky opponent can really ruin a PBEM game.

[Sheff]

I'm sure everybody will point their fingers and yell "herasy" for making such a mild suggestion in a sea of drastic ideas, but hear me out...

The idea is to limit the ability of a nation to rapidly reach a high tech level with a certain technology, right? So why not limit the amount of research points they can add each turn? By forcing a nation to only increase their assigned research in any given field by 1 per turn it will take 5 turns to reach maximum research. This way players have to plan in advance and their returns to their investments will, at the least, be slowed. You can also look at this as the building of a research project in the area (ie. start the project, increase funds, so-on and so-forth).

So bring on the criticism, I can't wait

[zappsweden]

quote of Sheff:

"So why not limit the amount of research points they can add each turn?"

That is also a good idea!

[ September 05, 2002, 07:16 AM: Message edited by: zappsweden ]

[John DiFool]

Originally posted by zappsweden:

quote of Sheff:

"So why not limit the amount of research points they can add each turn?"

That is also a good idea!

And is something which can be added to the game

system easily and painlessly.

JD

[arby]

Originally posted by Rediroc:

Well something should be done to limit the reseach accomplishable in one year. A very lucky opponent can really ruin a PBEM game.

I'm with you on that. I've been on both ends. In one game, my Brits had level 3 Ind Tech & Jets, and Lev 1 L/R Aircraft, by the end of 1940. In another game as allies, it's the middle of 1942 and I've had 1 -- count 'em, 1 -- tech advance among the three countries. My opponent, meanwhile, is at level 3 Jets, tanks, and armies, and level 4 anti-aircraft.

Here's a question: does anybody like the research system the way it is now, to the extent that they would prefer no changes be made?

[ September 05, 2002, 02:20 PM: Message edited by: arby ]

[Tha_Field_Marshall]

I like the random approach to tech. You nver know what will happen.

[USGrant]

I like the current system, I just think it needs a little calibration to synchronize the results with historical possibilities. In other words, Germany maxing out on the "important" techs by 1942 should be extremely rare, rather than the norm. A limit of 3 or 4 research points in any one area (rather than 5) should accomplish this without any purturbations to the system. It would also cause a little more variation in that the German (and later the allies) would have to spread out his points a little more to use them all, and therefore get advances in less common areas.

I also like the suggestion of one-per-turn assigned to an area.

[Brian Rock]

I also like the idea of limiting the points that can be added each turn. It means setting up an effective research program requires establishing infrastructure, training, and all the other stuff that goes with research IRL. Nice idea.

I prefer a system that has a combination of the predicitable and the random, but leaning towards the latter to simulate that nobody really knows that invest $X and you'll get a jet in between 12 and 15 months.

[Willard]

I think the issue of limiting research points is the key. However I believe the best way to accomplish this is that once the research is a success, you should not be able to reinvest that research point. For example, you buy one research point and invest it into industrial tech. When level one industrial tech is achieved, that research point is spent and cant be used again. That is the only way to simulate costs effectively, both time and money.

[USGrant]

Originally posted by Willard:

I think the issue of limiting research points is the key. However I believe the best way to accomplish this is that once the research is a success, you should not be able to reinvest that research point. For example, you buy one research point and invest it into industrial tech. When level one industrial tech is achieved, that research point is spent and cant be used again. That is the only way to simulate costs effectively, both time and money.

While the idea of expendable research points is a good one, I think implementation in SC is impractical. It would be a major curveball for the economic model. To retain the current, play-tested balance you would need to lower the cost of research points or increase the gross income available from the map. Either could work but could require significant rebalancing. For example, lowering the cost of research points could actually accelerate the early research - kind of opposite the intended effect!

With expendable research points the players could easily spend 6000 MPPs on research in a game, versus the current limit of 2500. Those additional MPPs come right out of armies, air fleets and submarines, hence the whole economic system must be rebalanced.

[Desert Dave]

As originally posted by John DiFool:

I have been carefully mulling over some alternatives

to the current tech system. Most people seem to

agree that it is too dependent on luck.

I guess I am in the other camp, the one with flimsy tent flaps and a piddling mud puddle instead of an Oasis, so OK -- why?

Luck arrives -- not willy-nilly and anthropomorphized as a fickle and capricious Greek God, but -- statistically, along a symmetrical "bell-curve," with MOST of the chance encounters bunched in the middle, and a gradually sloping curve until you reach, possibly, some outliers.

In other words, there will be very few games when one side gets all the advances, and the other only a few.

In the vast majority of games, each player will have roughly the same advances -- over the defined time span, as the other fellow. That's a statistical FACT, NOT the exception to the rule.

I favor the current set-up (... though, I wouldn't quibble about seeing a small limitation -- which could easily be a house-rule, such as no more than 3 or 4 investments in any one area) and am not put off by the occasional "outlier."

Let's take a hypothetical: Player A is of average tactical ability, whereas Player B is more experienced, and takes more time to complete each turn, and so is -- dedicated to success.

Player B gets slightly less (... remember this is the norm, not the exception) tech, while Player A, the less talented one, is smiled on voluptuously by ol' gamin-eyed Lady Luck (... though he may well be more Empirical and less Superstitious and so he will disdainfully disregard her whimsies, chalking it up to his OWN fast improving skill).

And so, the game is equalized. On any given day, like they say, yes? :cool:

Or, Player A gets LESS tech than hapless, unsmiled upon Player B. Now his troubles are magnified, and so he loses, and he loses his satin shirt and his gargoyle socks and the actual twinkle in his eye, alas.

So? If he is self protective, he says -- DADGUM IT! the ONLY reason I lost is 'cuz that other Cat got better tech! If he is disinterested in safeguarding his self-esteem, he says -- ho hum, that Evil Opponent got me this time, BUT, ruff-ruff, every dog has his day.

The point is that we NEED to have variability in a game that is so very limited in terms of what can be done over a 5-6 year span. You invade Poland, you do Scandinavia, you blitz the low countries, etc, etc, etc, etc.

So one game the other evil fellow got level 5 jets within 2 years and you, MOMENTARILY plucked naked and out of luck, got zilch -- what to do?

Well, surely you have gotten (or surely will get) a few advances of your own (... unless you are a social outcast or a statistical outlier) and must maximize what that particular advance or two has provided you -- an OPPORTUNITY, and NOT -- a reason to curse the screen and give up. You might switch a few points to Air, or Air Defense or -- change your Grand Strategy altogether!

Again, you must adjust and go with the flow. Rather than worrying about your particular place on the hiearchy (... begin beaten by what you consider a lesser opponent or lesser god, implies that you care to be higher or better than the status quo or the norm), you might learn from the occasion and eventually do better. We needn't win every time, do we?

To sum it up. I have no concern for the occasional strange game -- it can be instructive. Most of the encounters, however, WILL fall within narrow parameters given the nature of probability.

In any event, there are other factors always at work and play, and it may well have been one of those that caused all my troubles... or, all my success. In either case, and appreciating that a game is meant to provide fun & leisure, why worry about it?

[USGrant]

Originally posted by Brian Rock:

I also like the idea of limiting the points that can be added each turn. It means setting up an effective research program requires establishing infrastructure, training, and all the other stuff that goes with research IRL. Nice idea.

I prefer a system that has a combination of the predicitable and the random, but leaning towards the latter to simulate that nobody really knows that invest $X and you'll get a jet in between 12 and 15 months.

As I think about it, limiting the change in research points in an area to one-per-turn (plus or minus) would have a number of beneficial effects:

1. Slow down research a little bit

2. Encourage diversification of research, thereby making gamey results like Level 5 Air Fleets and Level 0 Bombers less common.

3. Cause players to think twice about rocketing to Level 5 using 5 research points because they could have 5 points assigned to a maxed area and could only move one per turn out (thereby wasting some research capacity for a few turns). This will encourage players to decrease research points in an area once they reach level 4, unless they really want level 5.

4. Slow down the end-game conversion of research points to armies.

[ September 06, 2002, 11:23 AM: Message edited by: USGrant ]

[USGrant]

Originally posted by Immer Etwas:

Luck arrives -- not willy-nilly and anthropomorphized as a fickle and capricious Greek God, but -- statistically, along a symmetrical "bell-curve," with MOST of the chance encounters bunched in the middle, and a gradually sloping curve until you reach, possibly, some outliers.

What he said!

[zappsweden]

Originally posted by Rediroc:

Well something should be done to limit the reseach accomplishable in one year. A very lucky opponent can really ruin a PBEM game.

By having a semi-random system the random spread for each advance will be lower and the probability for several quick advances will therefore be DRAMATICALLY reduced. Several quick advances almost works as a mathematical product. It will work similar to when comparing the probability to roll a die 6 compared to rolling three six's in a row.

[SeaMonkey]

Its like what Immer said. Its also your chance to beat that "cheatin" AI.