I don't think the "seed a ton of money and slow ride the interest wave to completion" procedure is the ultimate research method. Yes, you get great benefits by generating interest, but the seeding is problematic. By definition, a large seed deprives you of having earned interest on that seed. In fact, the larger the seed, the more inefficient your overall research spending will be. Instead, you will get the best results by starting slow at the beginning and riding the interest wave throughout the entire investment. The bottom line is that the essential idea of optimal research is growing your contribution at no more than the optimal rate. And using the "=" key is a great way to research near-optimally!
The research bonus is a somewhat confusing algorithm, so here is another way to think of it:
Each turn you can get as much as a tripling of your research spending (a 200% bonus added to your contribution) up to a cap. The cap is 15% of the cumulative amount you have banked for the discovery throughout all previous turns.
Also it is important to remember that the bonus is a cap, without a minimum requirement: there is no penalty for contributing less than the maximum you would be entitled to. It will just take more turns to get your discovery.
Anyway what this means is the best outcome is that when you hit the percentages for discovery, you only paid out 33% of the research cost, and you got 67% for free! This requires starting slow and, as mentioned before, growing your contributions at no more than the optimal rate.
Let's take an example and suppose you seeded 50% of the research cost. Right away, you have given up a ton of free money: (50% of the research cost * the 67% you have could gotten for free), or 33% of the total research cost. In effect seeding 50% slashes your potential bonus in half: you could have had a 67% cost reduction but instead you have a 33% reduction; you have gone from an optimal 200% spending bonus to an effective 100% one.
Another example: a 25% seed gives up (25% of the research cost * the 67% you could have gotten for free), or 16.7% of the total research cost. In effect seeding 25% slashes your potential bonus by one-quarter: you could have had a 67% cost reduction but now you have a 50% reduction; you have gone from an optimal 200% spending bonus to an effective 150% one.
On the other hand, now let's think about just hitting the "=" key to spread your research equally among all six fields equally. Over time you will tend towards the optimal rate of spending, and it is self-regulating!
For example, let's say early in the game you've built a large empire by focusing on expansion and factories, and deferred much spending on research. When you do turn to research in earnest, and start plowing money from fully-developed colonies into tech, your initial investments will be high relative to each discovery's research cost, so you won't be taking much advantage of the bonus. Another way to think of this is that you would be seeding high amounts, and on subsequent turns probably contributing in excess of the bonus cap. Not too efficient.
And yet! As you jump up to the next tier, your contributions are smaller relative to the increased costs of the more advanced techs. You are seeding a proportionally smaller amount, and you will be able to take more advantage of the bonus. Your research is automatically becoming more efficient.
The best case scenario will tend to come naturally by the mid-game: your initial seeds will be small, and your continuing contributions will take great advantage of the bonus. For example, let's say you're mid-game and spending only 1% of a hypothetical 10,000 RP discovery cost per turn. The potential interest lost to seeding is negligible. And without any bonuses, it would take you 100 turns to hit the percentages (1% * 100 = 100%); what a drag! But you're starting to enjoy the bonus very quickly; by turn 6 you have doubled your contribution (674 banked through turn 5 * 15% cap = 101 bonus), and by turn 9 you are tripling it (1,373 banked through turn 8 * 15% cap = 200 bonus). Now we're talking! And you hit the percentages on turn 37, instead of on turn 100. See the spoilered table at the bottom for the numbers for this example.
The bottom line is that the seeding method is an "in-between" method that can get you finish line of a discovery faster than steady contributions do, while taking reduced advantage of interest generation. For the long haul, though, it is wasteful, and you're probably better off overall spreading your research around. Even one tick, or a few, in every research category is maximally efficient.
(April 13th, 2016, 12:02)SvenBent Wrote: Wouldn't it put you in the same situation as the current MoO interest ?
for optimal purpose its better to pile in a bunch in the starting turns and ride the interest rather than doing a little investment reach turn. the only thing different is that no investment at all is required to gain interest in a turn.
Or maybe it's better just to focus all of your research on one tech at a time? I don't know. This would require discussion and play-testing, preferably during the alpha.
The one thing I don't want is for someone to conclude that "this is the BEST way to do it", because then all I've done is create a puzzle to solve instead of a decision to make.
Both. If there were no cumulative bonus, it would always be correct to research one tech at a time until each hit the percentages. But under a system where you get free RP proportional to your existing investment even when you spend nothing, optimal procedure - the "one right answer" - is to focus all your spending on one tech field for one (or a few if it's urgent) turns, then cease to spend on that tech until it comes in, while you focus all your spending on other tech fields and/or other priorities.
(April 13th, 2016, 06:50)Ray F Wrote: That's definitely better, but it still doesn't address my primary concern.
From a game design standpoint, I don't understand why we want to create a system that coerces a player to allocate his research spending in a particular way.
It's kind of like how MOO4 has diminishing returns for additional workers (4-3-3-2-2-1), basically forcing players to develop colonies in a way that the game developer thinks is best (balanced).
In that sense, I don't want to punish a player for stopping spending on a category any more than I want to reward him for stopping spending except for the specific number of ticks on the research bar. That's gamey.
There's a line you cross between needing to know the game elements (which tech you need the most, for example) and the game code (the minimum number of ticks to maximize research).
Because, it that sense, you might as well just make the minimum amount of ticks you are allowed to spend to be whatever gives you the max interest. Remember, if there's ever only one correct decision to be made in a situation, that decision needs to be removed from the game because it's not really a decision.
There is always going to be a theoretical optimum play for any given system, but in the research rules I described, unlike any of the others I've seen, this optimum is on a shallow slope and varies most significantly according to strategic objectives. It does not resemble MoO4's additional worker return because the research fields are small in number and non-interchangeable; putting extra points into Gatling Laser at a better rate-of-return per RP than Improved Eco Restoration is a strategic mistake if your priority is economic development.
The system I described does not intrinsically punish players for stopping spending; unless a separate penalty instituted, their existing investment will still be there whenever spending resumes. It is true that maintaining at least a small amount of spending in a tech into which you have already invested heavily would be more efficient than cutting off spending on that tech altogether, but this seems to me to be appropriate to the flavor of actually doing research, transparent (why would you want to suddenly stop researching a project in which you've already invested heavily, and why would you expect this to get you the tech as quickly or efficiently as continuing your investment?) and in keeping with the intended design of the original game.
That doesn't mean the quickly-calculated system I described is perfect, and I could well have missed something; it also is likely to result in faster research times (and/or more production available for other things) on average than in original Orion, as would the new-expenditure-independent interest system you described, to an even greater degree, but I think it does the job better than the other options I've seen.
(April 13th, 2016, 18:22)Zygot Wrote: On the other hand, now let's think about just hitting the "=" key to spread your research equally among all six fields equally. Over time you will tend towards the optimal rate of spending, and it is self-regulating!
ok, I had to separate this snippet out because, in my mind, this is exactly how it should work. I'll have to think for a bit on the rest of your post, but I wanted to make this part very clear.
Let me explain why "doing nothing" (pressing =) should be the "optimum" path for maximizing research points. Doesn't that sound unintuitive -- rewarding players for doing nothing???
Because, in reality, doing nothing means you have chosen to not manage the part of your empire that decides which of the 6 techs you need the most. Sure, sure, those Bulrathi transports are on the way, but we don't want to divert more research to our Plasma Rifles because that would be inefficient!
See where I'm going? The give and take of a competitive game will always pull players into trying to prioritize their spending so that they can get a particularly important tech a lot faster. That is the key advantage, a tactical one, of shifting spending around.
But pressing = and keeping spending even across all categories throws that advantage out the window in exchange for a different advantage, a strategic one, --- the most efficient use of research points over a long period.
That is the kind of choice that I think is interesting.
One more thing and then I'll shut up and let the real experts debate the relative merits and flaws of research allocation strategies.
There are 60 total "ticks" in the ROTP research allocation. Pressing '=' automatically assigns 10 ticks to each category -- an even distribution.
I can easily devise a research interest formula mechanic that works just like the original MOO1, except the maximum contribution with full interest "matching" is not 7.5% of the accumulated research but 10 ticks, or 1/6th of the total empire research.
This takes the "gamey" aspect out of it, which really bugs me, and rewards a balanced research strategy like the original game. But maintaining the optimum balance in all categories means that the player has to give up the advantage of prioritizing tech research.
(April 13th, 2016, 19:54)Ray F Wrote: One more thing and then I'll shut up and let the real experts debate the relative merits and flaws of research allocation strategies.
There are 60 total "ticks" in the ROTP research allocation. Pressing '=' automatically assigns 10 ticks to each category -- an even distribution.
I can easily devise a research interest formula mechanic that works just like the original MOO1, except the maximum contribution with full interest "matching" is not 7.5% of the accumulated research but 10 ticks, or 1/6th of the total empire research.
This takes the "gamey" aspect out of it, which really bugs me, and rewards a balanced research strategy like the original game. But maintaining the optimum balance in all categories means that the player has to give up the advantage of prioritizing tech research.
Wow ... that's a really intriguing proposal, actually. Off the top of my head, it makes very-early research a lot less attractive (because a lot more research would be "wasted" either by spending in non-bonus territory or on fields you don't yet need) and mid-game changes in research allocation very rare (used mostly for emergencies) - just because a +200% research bonus is so powerful.
My instinct is that this would result in a less-interesting game overall though, as it nearly removes the choice, for most of the game, of which tech field to pursue: You'll never want to get off of "=" unless one of the techs that you're researching is at least three times as valuable to you as at least one of the others. That can be adjusted though, obviously:
1) You could reduce the value of the bonus (and presumably reduce the costs of techs to compensate). At +50%, for instance, or even (to a lesser extent) +100%, trading some portion of the bonus for a more-critical tech would be costly but not so much so as to virtually prevent all mid-game deviation from "=". In this case, standard practice would probably be to hit = and then respond to certain needs by allocating all ten points from one (or more) field(s) to a more critical one.
2) You could use a version of the slope idea I mentioned, but recalibrated to slider clicks; for instance, for a total close to the total bonus available in the base game, the first click in each field could give a bonus of +360%, with each subsequent click getting a smaller bonus until the tenth gives just +36% (and 11th and later clicks give no bonus at all). This would make abandoning a field completely very costly in terms of efficiency, and diverting from "=" would still be inefficient, but the difference around the margins would be small enough that meaningful tech spending allocation choices could be made a lot more often without crippling the empire's overall research entirely. The problem is that this would only be true around the margins: Doubling the research speed for any given tech would be no better (in fact, very slightly worse) in this case than in the straight "+200% for the first ten clicks" proposal.
3) Combine #s 1 & 2. Perhaps reduce research costs by roughly 30% and grant the first click in any field a +250% research bonus, dropping evenly from there, down to +25% for the tenth click and +0% thereafter. That ... actually, I think that could be pretty cool!
(April 13th, 2016, 21:40)RefSteel Wrote: Wow ... that's a really intriguing proposal, actually. Off the top of my head, it makes very-early research a lot less attractive (because a lot more research would be "wasted" either by spending in non-bonus territory or on fields you don't yet need) and mid-game changes in research allocation very rare (used mostly for emergencies) - just because a +200% research bonus is so powerful.
I don't want to get too hung up on the amount of the research bonus; that can be tweaked obviously. +200% is way too much of a bonus for not focusing on techs.
The point is there should be a bonus of sorts just to create the decision between focus and balanced. I'm ok with that. You just can't realistically do both; the ticks involved to maximum the research bonus need to be enough to make focusing difficult.
I'm just not ok with the player having to regularly twiddle sliders to get the right bonus and to constantly keep going back to adjust.
The OSG currently states:
Quote:The maximum interest that can be earned is 15% of the amount presently invested in researching that item, but never more than twice present BC investment in that technology this turn
Maybe change it to something like:
Quote:The maximum interest that can be earned is 50% of the present BC investment in that technology this turn, but never more than 1/12th of the empire's total research this turn. The interest is applied as RP in the subsequent turn.
This means that, if you press '=' and do nothing else, you will get a 50% research bonus across all categories. But if you focus and spend more than 10 ticks in a category, you get no additional interest *and* you lose some interest in the tech categories that you deducted from.
(April 13th, 2016, 23:19)Ray F Wrote: I don't want to get too hung up on the amount of the research bonus; that can be tweaked obviously. +200% is way too much of a bonus for not focusing on techs.
Yes. Also, after further consideration, I realized that of the three options I listed, the first (just tweaking the research bonus as you say) was best. It isn't important to be able to fiddle around on the margins; it's important to make big, important calls.
Quote:The maximum interest that can be earned is 50% of the present BC investment in that technology this turn, but never more than 1/12th of the empire's total research this turn. The interest is applied as RP in the subsequent turn.
Yeah, that sounds right. Tech costs would probably have to be adjusted though, I suspect. Maybe even cut in half. That would have unintended consequences too, but maybe only deep in the tech tree where most games never go....
(April 13th, 2016, 23:39)RefSteel Wrote: Yeah, that sounds right. Tech costs would probably have to be adjusted though, I suspect. Maybe even cut in half. That would have unintended consequences too, but maybe only deep in the tech tree where most games never go....
Well, the other shoe to drop is that tech costs need to scale in some way based on galaxy size. MOO1 did not have extremely large maps (100 systems was the largest) so it did not have to concern itself too much with this problem. It instead scaled tech costs based on the difficulty level. However, on very large maps you don't want a situation where the player is already halfway through his tech tree before he encounters another race.
My point is that there is already going to be a separate knob for overall tech costs, so those will need to be tuned as well. These things will hopefully wash out in the alpha.
Thrawn, I don't think the existing formula can be tweaked so easily. Any formula will lead to some sort of optimization and slider-tweaking each turn.
I am convinced now that the way to go is to make the "optimum" interest value require so many ticks (1/6th of total) that it is impossible for a player to choose both paths of maximizing RP interest or having the ability to focus spending in the tech categories. In other words, he has to make that "interesting decision" between long-term strategic planning (maximizing interest) or short-term tactical responses (focusing RP spending). He can't do both. That's what changes it from a puzzle (with a best choice) to a game (with competing choices).
And by using 1/6th as the mark, the "=" button removes all of the tweaking.
For example, let's say that you want to optimize RP interest in Planetolgoy and Force Fields, two tech categories that perhaps do not have immediate tactical interest. But you still want to be able to focus research on the other 4 categories.
All you'd have to do is press "=" and then LOCK the Planetology and Force Field categories. Done! Now you are free to focus spending in the other 4 categories as your tactical needs require WITHOUT ever screwing up the interest optimization in the locked categories. You easily get the exact amount of control that you want with zero micro-management!
Honestly, the more and more I think about this, the more elegant and intuitive it seems. All of the intellectual discussion that you experts are engaging in has really helped me stumble onto a beautiful solution, imo.
I really like your 1/6th of the slider solution, Ray F! Brilliant!
I agree that something like a 50% bonus up to the 1/6th mark would be reasonable. It is tangible enough that it will noticeably hurt to give it up, and yet it is not so high as to completely dissuade a player from ever focusing on some techs disproportionately for tactical reasons.
