Research

[Briggs]

Are we limited to one tank advancement per year? I've dumped full (5) research points into Heavy Tanks as Germans yet can only still pull off a level 2 heavy tank. I've also only been able to get to level 1 anti-tank weapons, level 2 Industial Tech, and level 2 radar. Is there a limiter, or do I just have bad luck?

[SuperTed]

PL,

If you go here, you can read this:

"...Each research type has a maximum level of 5 and a maximum number of 5 research points that can be put against it. At the beginning of each turn, each research point has a 5% chance of returning the next level of research (so 5 research points in anti-tank research will give you a 25% chance per turn of getting the next level), so the more research points put against a particular type of research, the better the chances of achieving the next level..."

[pzgndr]

Research *seems* slow to me also. I can usually have Germany up to 4 research points by August after France falls, maybe 5 later. In my many games, I usually see only 2 technology advances, sometimes 3, and once got 4. Statistically there should be more, at least occassionally on a particularly lucky day. I had initial concerns about some of the fantastic research advances some players reported, but I haven't seen them, at least not on normal settings and above.

I assume research is discussed in more detail in the manual, specifically the design decisions for how many research points and what tech levels each country starts with for each scenario. It would be interesting to track where each country should be historically, and how well the average player can reproduce those results - from affording the research points to seeing advances based on average chances.

Based only on the demo, I can't tell if I'm expecting too much or doing just fine. Time will tell. We can always edit the scenarios if we're still uncomfortable with the long-term research results. Frankly, I think every country should start with at least 1 free point in 1939, and Britain and Germany with 2 in 1940, just to generate some surprises if nothing else. The whole research thing and various strategies to pursue will be an excellent topic for further discussion later on once we get some campaign games under our belts.

[Desert Dave]

As originally posted by Bill Macon:

Research *seems* slow to me also.

I recently had a game where Italy was up to Industrial level 2, Jet-Air level 1, and Gun Laying Radar level 1, all by demo end.

And once, having invested in 5 separate categories for the Germans -- got ALL FIVE on one and the same turn, I kid you not!

To the larger point, I have always argued that Research decisions and advancements would have a VERY SIGNIFICANT impact on each game played, and have applauded and appreciated that game-design decision. :cool:

In fact, these random (though, you must invest first, so it is not so fatalistic at all) advancements would actually DRIVE future tactical decisions, in that -- let's say you had invested a point in rockets, and by early 1943 you are up to level 4... wouldn't you THEN decide to deploy a raft of rockets along the French north coast?

This is surely a topic that should, and will be discussed further, once we gain more full-length game experience. Indeed, IMO, it is an extremely critical aspect of the game.

[Guest Mike]

There's something not right about research tho.

The Sov's have arguably the largest ability to perform research given their huge MPP potential, yet in "real life" they had a completely minimal capacity to do any real research.

The recent screenshot by someone of a huge Sov army with a couple of rocket units in the back line bought this home to me in a big way!!

Now don't get me wrong - I think the game is pretty cool, but it doesn't seem to be any great advance on "Hitlers War", which I enjoyed greatly. In fact I'd go so far as to say that it's significantly lacking in some areas compared to HW (although it's ahead of it in others for sure).

Basically we seem to have a very simple game translated onto the computer. It doesn't really use the computer's mathematical ability to any great degree (excepting the AI I'd guess) to give us simple presentations of complicated mechanisms.

Of course it also comes fairly cheap, which is a major bonus!! lol

Overall I think SC rates about a B or a B-. It's not a bad game at all, and I shall enjoy playing it a lot.

But it fails to live up to the potential of today's computing power.

[Hubert Cater]

Research *seems* slow to me also.

With the demo, you are only playing a single year of a possible seven

Hubert

[Silvercloud]

It seems to me that for a possible future release (SC2) there would be a need to specify research by country, allowing for faster research than originally happened in the war (also by Russia), but culminating in the type of units seen in panzer general, this would balance the research effort, as if the research of the Russians is likely to come up with less radical advances than the Germans in plane tech, but equal in tank tech and slightly better in industrial tech. However this might require a rebalancing of the whole game (MPP wise and such).

Definitely a long term goal...

[pzgndr]

Originally posted by Hubert Cater:

With the demo, you are only playing a single year of a possible seven

Yes, I do acknowledge that. My point was that with 4 research points by early August and about 17-18 turns left in the demo, the Germans should statistically see 3-4 advances on average. Sometimes more, sometimes less. Seeing only 2 advances more times than not indicates I'm consistently unlucky or something else may be amiss. After reading IE's post, I decided to just spread everything out and damn if I didn't see four advances all in the same turn. That's a .000625% chance, for at least two of us! I got 1 more during that game, so now my best is 5.

I'm not questioning that these events can occur, simply that the averages don't *seem* right. Each turn's research determination should be completely random and independant of other research results, and perhaps there's a minor bug somewhere that skews things. So just looking for assurance that the code is perfectly OK and that these skewed events are purely the result of Lady Luck.

[Desert Dave]

As originally posted by Bill Macon:

I'm not questioning that these events can occur, simply that the averages don't *seem* right. Each turn's research determination should be completely random and independant of other research results, and perhaps there's a minor bug somewhere that skews things. So just looking for assurance that the code is perfectly OK and that these skewed events are purely the result of Lady Luck.

You know, now that I think about it, I was sort of bothered by these kinds of success probabilities as well -- it's just that the very fortunate results furthered my private war aims so I kind of... let it slide in my mind.

For the number of games that I have played, this kind of result -- where 4 or more research advances occur on one turn -- has happened at least 4 times that I can remember!

If Bill's % calculations are correct (and I must accept, since my Statistics Intellect is somewhat derelict), then this kind of result does seem skewed, almost as if the random-number generator gets STUCK? :eek:

I don't want to fall from Science-grace and get too weird -- by believing that the program is somehow getting excited for my success, and therefore -- WILLINGLY HELPING ME, but I agree with Bill that this might be something that needs looking into?

[Hubert Cater]

OK I'll take a look at this as well

Hubert

[Bruce70]

Just to give something concrete to go on...

Here is what you should expect for 4 research points over 17 turns:

0 advances - 2% of the time

1 advance - 10%

2 - 19%

3 - 24%

4 - 21%

>4 - 24%

And just to add something else to the mix. I played on the lowest level as the Germans (and disbanded most of my corps) so that I could experiment with research and on two occasions invested 5 points in industrial tech and 5 in tanks. On both occasions I received 3 tech levels of industrial tech but 0 of tanks.

The chance of seeing 0 advances over 30 odd turns (2 games) with 5 points invested is 0.02% So maybe research is being calculated differently for different techs???

While anything is possible it does seem that something weird is going on here.

[Briggs]

One thing you guys should be aware of if you have not already noticed, is that tech makes your stuff EXPENSIVE. A German sub (Tech level 1) costs 358. An Italian sub (Tech level 0) costs 325. Just as a heads up to you guys, you might want to consider bringing in your Industrial Technology first, then using heavy tanks, etc.

Although, have you ever rolled into Russia with Heavy Tanks 2? It's a blast! Can't wait for the whole game.

[Bruce70]

The extra cost only reflects the extra unit strength gained (if applicable). eg at level 1 a sub has a max strength of 11 points compared to 10 for a level 0 sub and 358 is 10% more than 325

You usually get combat bonuses in addition to the max strength bonus so the research is worth it even without industrial tech. But I agree with you, industrial tech is nice.

[Bruce70]

I played a few games last night. Lowest difficulty, capture low countries in one turn, then the next turn disband enough units to give me 2000 MPPs to put into research.

By spending all these points immediately I have 8 points invested for 14 turns. Note that there are only 14 turns remaining in the demo not 17 as was stated earlier. Using a variety of investment strategies I was not able to produce any unusual results. I guess you only pay attention to the research when something unusual happens.

The 14 turn limit changes the table I produced earlier to the following expectations with 4 points invested:

advances / %chance

0 / 4%

1 / 15%

2 / 25%

3 / 25%

4 / 17%

>4 / 14%

So there is about a 50% chance you will get 3 or more advances, which makes Bills experiences seem a little more acceptable but still unlucky enough to make him an excellent PBEM partner.

[ July 23, 2002, 07:53 PM: Message edited by: Bruce70 ]

[pzgndr]

These numbers look fine and I agree the overall results I've seen are within the reasonable bounds. I said *seemed* because my unofficial results tend to be on the low side of average, although I have had a few lucky games. Also, the fact that IE and I have seen the statistically rare 4-advances-all-in-the-same-turn event is cause enough to raise an eyebrow. (Then again, maybe we're playing the demo so much we're expected to see this! :eek: ) Perhaps there's some random number cache that isn't being refreshed during a turn's research checks or something. If it's anything at all, then it's probably very subtle since it hasn't been a concern for most players. Our good mechanic Hubert said he'd take a look under the hood and check out the funny noise, and we can't ask for more than that.

[pzgndr]

still unlucky enough to make him an excellent PBEM partner

(Sing along, Hee-Haw fans)

Gloom, despair, and agony on me

Deep dark depression, excessive misery

If it weren't for bad luck, I'd have no luck at all

Gloom, depair, and agony on me

[Gregory Barnard]

Las Vegas makes its living from people who think that chance plays fair, evenly, across a random series of attempts. We all know this is not the case. A small number of tries (<20) should not expect to return an average outcome in all cases. Etc, etc. Unless there is an issue with the random number generator seed, I'm sure that expectations should be weighed across the number of attempts across games. Perhaps a look at historical technological advances is appropriate. A couple good die rolls on Research should NOT dramatically influence the outcome of the war -- although I agree they could significantly influence strategy, etc.

[Bruce70]

Las Vegas makes its living from people who think that chance plays fair, evenly, across a random series of attempts.

Actually Las Vegas makes its living from the fact that chance does play fair - given enough trials. But you are right, the few strange results that have been experienced are nothing to worry about unless they can be consistently reproduced.

BTW I have never experienced 2 or more advances in one turn, has anybody else (apart from those already mentioned)?

[Briggs]

I would have to say I don't exactly agree with you there. I feel that the weight of the technologicaly superior WILL make their presence felt. I have gotten a level 5 Heavy Armour unit for the Germans just to see what would happen, and I must say I made rather short work of the Russians rather quickly. Also, long range fighters prevent the need to rebase as often, so you don't continualy have to halt barbarossa.

[Scott Udell]

In my one full game ('39-'46) that I've completed so far, research "felt" too slow to me as well (note the quotes--I haven't done any statistical analysis on this yet). I did manage to get the Americans to 5 points in industrial tech, and fairly quickly, but only with five research points applied. Other techs advanced only very slowly, and one (for the Americans--anti-tank) not at all.

Now, unless I'm misunderstanding things (or have just been really unlucky), then I think a change might be in order: keep your random chance each turn, but each turn you dedicate a point you get that point's worth of percentage (or half, or one third, or one quarter) as a starting point next turn. So, one turn, one point, 5% chance; next turn, one point, you get 2.5% of "groundwork" chance, then 5% for the point, for a 7% chance. Then next turn 5% groundwork, 5% for the point, for a 10% chance. If you take your all your points away, research stops, but if you apply it again later, your "groundwork" is still there. I don't know if this would make sense across the board, and maybe one side should be better in one area than another (although I think the mechanic of giving sides research levels at the start of a scenario to reflect advantage works too).

[J Wagner]

Hey Scott...since you come here I have to ask...has the CG forums been scrapped?....my CG bookmark takes me to a new setup w/o the forums....sorry to digress off the topic everyone..

[gorski]

Las Vegas makes its living from people who think that chance plays fair, evenly, across a random series of attempts

I know bruce already commented on this, but I had to throw my 2 cents in. The casinos make their money by having a better chance of winning than you do on each bet.(blackjack, craps, etc) If you count cards at blackjack and play with the right rules, you turn the tables on the casino and have a slightly better chance of winning each bet than they do. If they catch you card counting, they will kick you out(or worse ) because they are a business not a charity.

Gorski

[Aloid]

Originally posted by Scott Udell:

<snip>... (ouch)

Now, unless I'm misunderstanding things (or have just been really unlucky), then I think a change might be in order: keep your random chance each turn, but each turn you dedicate a point you get that point's worth of percentage (or half, or one third, or one quarter) as a starting point next turn. So, one turn, one point, 5% chance; next turn, one point, you get 2.5% of "groundwork" chance, then 5% for the point, for a 7% chance. Then next turn 5% groundwork, 5% for the point, for a 10% chance. If you take your all your points away, research stops, but if you apply it again later, your "groundwork" is still there. I don't know if this would make sense across the board, and maybe one side should be better in one area than another (although I think the mechanic of giving sides research levels at the start of a scenario to reflect advantage works too).

This is a nice way to reflect accumulated research over time. I'm wondering if SC should also require continued investment each turn to go along with it.

Right now, the current process works for me. It gives a different out come each game, which will be needed to help extend the lifespan of this game.

It would be a nice thing to enhance for a patch or SC2.

Aloid

I really think I beat out Alley Cat for that first order...

[pzgndr]

The current system is fine, without additional baseloading to lead to a definite result. We just need to ensure each individual research check is in fact random and mutually exclusive of all other events. Odds being what they are, I like the idea of surprises each game for replayability. My concerns are not a result of any sour grapes or anything, only a gut feeling that something may need a tweak.

One enhancement I would like to see would be some sharing of research advances among allies, since this would be expected to some degree. (Russia may be a different story.) One way to do this may be to assign a virtual research point to areas where an ally has already made a gain. For example, Germany with level 1 subs would give Italy 1 virtual point for sub research, a point which could not be used elsewhere and would go away once parity was achieved. If 5%/point is too much for this and affects play balance too much, then something less for virtual points could be considered. Just an idea.

[DevilDog]

Usually I buy 5 points and put them into one catagory. But I probably average only one advance per game. My one game exception to this supports the "there's something wrong" theory. I had one turn where I put 5 points into Industrial Tech and the next turn I had an advance. And the next. And the next. And the next. And the next. I ended up with Industrial Tech maxed out in just five turns. It was like the die roll got stuck. I have never been able to replicate this though.