VVV scoring system suggestion - Updated

[sanabas]

sanabas, I don't know if I will purchase BNW or not but this is how I see the future with your suggested scoring system.

Making HUGE maps more valuable than standard could be a big mistake.
Most people don't have computers capable of playing huge.

My PC is at least 5 years old, and was only mid-range even then. It plays huge games ok. Can take some time between turns, but I can still play them.

Marathon games should not be more valuable than standard speed.
Most people don't have the time to play them, some people would find them tedious and others would find them to be not much of a challenge.

Just suppose that a stupid old git like me was to play, how would I approach it?
I would just rack up a load of huge marathon deity games, and these are the reasons why.

1 - I would buy a new pc.
2 - I have forty hrs a week spare time.
3 - I am not capable of playing Deity games in the true sense so this would be my cop out.

Get to it.

The ELITE players of the game are all agreed that standard speed is the most correct speed for a challenge, so why not bracket standard speed alone as the most rewarding to play. Quick, Epic and Marathon can then be classed together as less rewarding points wise! Personally, I would just make them all equal, I can just imagine all the moaning and back biting that might go on due to the fact that a stupid old git like me has lots of time to abuse the scoring system.

I hope this helps with your last minute tweaking.

Good luck and good gaming to all.

Again, these are points for empty tables. Sure, you can attempt to abuse the system by playing lots of huge/marathon/deity games. You'll get ~250 points per game. But if standard speed, standard size is more popular, say there are 4 games on a std/std/deity table, whoever is in first will be getting ~300 points.

This is here as an incentive to fill empty tables, as a reward for playing games that likely won't get much comp. Instead of getting nothing, instead of waiting months or years for somebody else to try and fail to beat your score, giving you a bronze medal, you'll actually get something. There's more incentive for the harder/bigger/longer games, because they're the ones that are far less likely to see others attempt them. Just 3 people play the same std/std deity game, and 1st is worth more than an empty huge/deity table. Beat somebody else's std/emperor time, you'll get ~120 points. Set your own time on an empty huge/emperor, you'll get ~120 points.

And regardless of how points get assigned, it's pretty much guaranteed that it will be possible for someone with lots of time, who doesn't care about anything other than their VVV score, to pick the most effective way of playing 'cop out' games to get points. We can either be very restrictive, both in settings allowed and in counting only a small number of games, or we can be more inclusive, reward volume to some extent. If we're restricitive, then participation will likely drop, players will feel forced into certain leaders, certain game types, or players will already have the maximum achievable score (look at civ3, I haven't played that in 7 years, and I'd still only need to play 3 games to get back to the maximum possible score in 3 of the 4 quartermaster categories). If we're inclusive, then those who spend a lot of time playing will get a boost due to sheer volume. This is an attempt to get a balance between those things, to avoid people feeling like they need to play the same game over and over because that's where the easy gold medals are, to avoid feeling like you have to keep rerolling as the inca (or spain for civ5) because that's the only way to challenge the top times, to avoid feeling like most games are pointless because they'll never be worth even a bronze, etc. If you feel you can abuse this setup, then please do.

 

[TractorBoy]

My PC is at least 5 years old, and was only mid-range even then. It plays huge games ok. Can take some time between turns, but I can still play them.

Lots of people (including me) have complained in the past that their pc could not handle huge games, maybe it was something to do with game design, I got around the problem back in July 2012 by purchasing a new pc. Maybe things have improved for all concerned and if this is the case then my suggestion becomes irrelevant and nonsensical.

I don't use computers in the same way that most people do, I just see them as a game console, if the only way to play poker or Civ5 was through a play station then I wouldn't even own a pc.

 

[TractorBoy]

Get to it.

Me quoting "I" does not necessarily mean me, more a case of where others might look for quick points. If I do purchase BNW, I just know that I would become committed and that is exactly what worries me for my own peace of mind and consequently my family relations!! I need to show some will power and give up this game completely.

 

[sanabas]

Yeah, not sure. I've never had a problem, but I also think I have my graphical settings lower than average, especially going off other people's screenies. My water doesn't look like theirs.

But even if half the players simply can't play huge, other than 1 game to qualify for VVV, I still don't think it will matter. There will be plenty of points on offer for non-huge games. 6 months in, hopefully with a lot of active players, and I expect the highest available scores won't be from empty huge games, they will actually be from more competitive standard games. I don't think people will feel forced to play a lot of huge maps. But equally, I don't think people will feel forced to play duel maps, or to ignore huge maps, which is certainly something I've felt when playing G&K.

Me quoting "I" does not necessarily mean me, more a case of where others might look for quick points. If I do purchase BNW, I just know that I would become committed and that is exactly what worries me for my own peace of mind and consequently my family relations!! I need to show some will power and give up this game completely.

Marathon huge deity games may be easy points, but they certainly won't be quick points.

 

[TractorBoy]

This is here as an incentive to fill empty tables, as a reward for playing games that likely won't get much comp. Instead of getting nothing, instead of waiting months or years for somebody else to try and fail to beat your score, giving you a bronze medal, you'll actually get something. There's more incentive for the harder/bigger/longer games, because they're the ones that are far less likely to see others attempt them.

Just in case you hadn't noticed, I was the first person to bring this point to light on the forum, and I have talked about it over and over. I was also the first person to suggest abolishing medals and reward participants down to tenth place.

And regardless of how points get assigned, it's pretty much guaranteed that it will be possible for someone with lots of time, who doesn't care about anything other than their VVV score, to pick the most effective way of playing 'cop out' games to get points.

Ha Ha!!, I wouldn't call Marathon Deity cop out games, they are just below the very best standard that I achieved playing VANILLA. My statement of cop out is much more a case of how it would be viewed by some of the top guys. You have just proved my assumption to be correct. LOL.

We can either be very restrictive, both in settings allowed and in counting only a small number of games, or we can be more inclusive, reward volume to some extent. If we're restricitive, then participation will likely drop, players will feel forced into certain leaders, certain game types, or players will already have the maximum achievable score (look at civ3, I haven't played that in 7 years, and I'd still only need to play 3 games to get back to the maximum possible score in 3 of the 4 quartermaster categories). If we're inclusive, then those who spend a lot of time playing will get a boost due to sheer volume. This is an attempt to get a balance between those things, to avoid people feeling like they need to play the same game over and over because that's where the easy gold medals are, to avoid feeling like you have to keep rerolling as the inca (or spain for civ5) because that's the only way to challenge the top times, to avoid feeling like most games are pointless because they'll never be worth even a bronze, etc. If you feel you can abuse this setup, then please do.

I fully understand and appreciate everything you have written here, and what's more I concur with most of it. What I don't agree with is your assumption that all of us play the same game over and over just for the sake of VVV score. Some of us are not as clever as you and it takes us a long time to hone our skills in a particular area, as a result we have to play over and over in order to reach somewhere near perfection. I have been playing duel deity for some time now and I still don't think I have quite worked it out properly.

 

[sanabas]

Ha Ha!!, I wouldn't call Marathon Deity cop out games, they are just below the very best standard that I achieved playing VANILLA. My statement of cop out is much more a case of how it would be viewed by some of the top guys. You have just proved my assumption to be correct. LOL.



I fully understand and appreciate everything you have written here, and what's more I concur with most of it. What I don't agree with is your assumption that all of us play the same game over and over just for the sake of VVV score. Some of us are not as clever as you and it takes us a long time to hone our skills in a particular area, as a result we have to play over and over in order to reach somewhere near perfection. I have been playing duel deity for some time now and I still don't think I have quite worked it out properly.

You did call marathon deity cop out games. I'm only using 'cop out' and playing just for the sake of VVV score because of what you posted:

Just suppose that a stupid old git like me was to play, how would I approach it?
I would just rack up a load of huge marathon deity games, and these are the reasons why.

1 - I would buy a new pc.
2 - I have forty hrs a week spare time.
3 - I am not capable of playing Deity games in the true sense so this would be my cop out.

If you're playing duel deity games now mostly to try and perfect your ability to play duel deity games, then why would you change your approach so radically if you got BNW? Why wouldn't you continue to play duel deity games over & over, perfect them with the more complex game of BNW?

Your original post tonight basically says that if you got BNW, you would play the self described cop out of marathon deity, because it's easier and you're not quite up to 'true' deity level, and because that's how you'd rack up a really big overall score.

How else am I meant to take that?

 

[TractorBoy]

You did call marathon deity cop out games. I'm only using 'cop out' and playing just for the sake of VVV score because of what you posted:
If you're playing duel deity games now mostly to try and perfect your ability to play duel deity games, then why would you change your approach so radically if you got BNW? Why wouldn't you continue to play duel deity games over & over, perfect them with the more complex game of BNW?

Your original post tonight basically says that if you got BNW, you would play the self described cop out of marathon deity, because it's easier and you're not quite up to 'true' deity level, and because that's how you'd rack up a really big overall score.

How else am I meant to take that?

Sorry, yes it does read that way, but was not my intention. I have just become used to people inferring that I am on a cop out. I call myself a stupid old git because of the way I have been spoken to on this forum and the general forum.

I would have thought the reason was obvious for a person of your intelligence. I play duel Deity games because they are the only ones that any number of people have competed for. I have since moved on to the tiny Deity games, before that I played an epic standard size domination deity game. I have spent 200 hrs trying to win a standard standard domination Deity game which has just one entry. I will complete it even if it takes me three years of trying such is my resolve. I'm not bothered whether I beat the other guy or not, I just want to achieve what I am finding very difficult.

My original post tonight says exactly what you say it does, but I never jump in at the deep end, I would have to find out how to play the game first. I would imagine some of the guys will do exactly what I said to get an early VVV lead.

You can take it how you like.

 

[Denniz]

I am working on understanding sanabas' scoring system and converting it to code. For this update, I still have some basic code to complete just to get the update done. So please be patient with me.

 

[beezany]

I like sanabas's proposal, especially the sliding scale for filling empty tables and the “option B” system for game diversity. However, I think the diversity system needs a little tweaking.

If it was aiming for 100 games per category, that's be:

LoN - 3 per nation (102)
Mapquest - 5 per map (100)
Tempi - 25 per speed (100)
Distance - 17 per mapsize (102)
Machiavelli - 20 per VC (100)
Inferno - 13 per difficulty (104)

While I like the basic idea of this system, I'm concerned about the high number of games per setting for Distance, Machiavelli, and Tempi. That lets people rack up a lot of quick and easy games before the limits ever affect them. Meanwhile, the people who want to complete all 100ish games will need to slog through 17 Huge maps and (even worse) 20 Time victories. Thus, this seems to reward gaming the system and penalize completists.

The “option C” averaging system helps to address the quick & easy half of the problem, but it still has the slogging half of the problem – perhaps even more so, because at least “option B” lets you completely ignore them until the end.

I would prefer an “option D” that encourages people to play on all settings without forcing them to play a huge fraction on annoying settings. I would like an approach that allows for some flexibility, so that you can play the majority of games on your favorite settings, and you can minimize one or two hated settings, without going too far one way or the other.

 

[beezany]

I may have found a good “option D” for calculating game diversity. The logistic function is an S-shaped curve that grows slowly from 0, increases rapidly through ½ at the midpoint, and then slows again as it approaches 1. You can apply the logistic curve to category completeness to reward diversity in a flexible way. It should be easy to calculate in most programming languages; the only tricky part is getting the scaling right.

The basic function: S(x) = 1 ÷ (1 + exp(–x))

How to use it: Let's take “Go the Distance” as an example. There are 6 different map sizes. If you submit one game for each size, you get full points. Otherwise, multiply your scores by S(+2) for 5 different map sizes, S(+1) for 4 sizes, down to S(–2) if you only use a single map size. That gives you a ⨉0.12 score multiplier for one map size, ⨉0.27 for two sizes, ⨉0.5 for three, ⨉0.73 for four, and ⨉0.88 for five map sizes.

I would apply this to games in sets. First, take the player's best game from each map size. If it's a complete set, they get full score, otherwise apply the logistic function as above. Then take another set from the next-best games. If it's complete set, they get full score, otherwise it gets curved. Keep going until you run out of games.

For example: Suppose that you submit 5 duel maps, 2 tiny and 2 standard maps, and 1 each of small, large, and huge. Take the best game from each size. It's a full set of 6, so the player gets full points for those games. Now take the second-best games from duel, tiny, and standard. That's only three map sizes, so the player only gets ⨉0.5 score for these games. Finally, take the remaining duel games. These all get a ⨉0.12 score multiplier because there's only one size left.

With this system, if you predominately play just one map size, you'll keep getting points, but only a few. Conversely, if you mostly skip one size, you'll lose some points, but only a few. (In fact, it's the same number of points, because the logistic function is symmetric.)

You can apply this similarly to the other categories; they'd just need to be scaled differently to suit the different numbers of settings. And you can tweak the map size numbers above too, if you want a gentler or harsher curve.

 

[beezany]

Let's run a couple of scenarios comparing options B, C, and D for “Go the Distance,” one of the categories most likely to be affected by the different methods. First, a recap of the scoring rules.

Option B: Total score is based on your best games for each map size, with no more than 17 games counted for each size (100÷6).

Option C: Calculate the average score for the map sizes (total÷6). If any single size exceeds this limit, the excess points are lost.

Option D: Collect games into sets with only one of each map size. Complete sets receive full score. Incomplete sets have their score reduced by the logistic function as described in my previous post.

Now let's compare three different players who have each submitted 24 games. To keep things simple and focus on the scoring options, let's assume that each game is worth 100 points.

Player 1 takes game diversity very seriously. He plays exactly 4 games for each map size.

Player 2 hates Huge games. He prefers Standard size for regular play and Duel for filling out VVV requirements, but otherwise he plays the map sizes pretty evenly. He plays 6 Duel, 4 Tiny, 4 Small, 6 Standard, 3 Large, and 1 Huge.

Player 3 only plays Duel games and Gauntlets. He has 17 Duel, 1 Tiny, 1 Small, 2 Standard, 2 Large, and 1 Huge.

How do the three players score in the different systems?

Option B: All three players are under the 17-game limit, so they all score a full 2,400 points.

Option C: No player can score more than 2,400÷6 = 400 points per map size. Player 1 still gets 2,400 points, player 2 gets 2,000 points, and player 3 gets only 1,100 points.

Option D: Player 1 still gets 2,400 points. Player 2 has a complete set (600), two sets of five (880), a set of four (292), and two sets of two (108) for 1,880 total points. Player three has a complete set (600), a set of three (150), and fifteen sets of one (180) for 930 total points.

What happens if each player adds another Duel game for VVV progress and another Standard game for a Gauntlet?

Option B: Players 1 and 2 are still under the cap, so they score a full 2,600 points. Player 3 is one game over the cap for Duel, so he only scores 2,500.

Option C: The new cap is 2,600÷6 = 433 points per map size. Both games put player 1 over the cap, so he only gets 66 points for 2,466 total. Player 2 was already over the cap for those sizes, but now the cap is higher, so he also gets 66 points for 2,066 total. Player 3 is still under the cap for Standard, so he actually gets 133 points for 1,233 total. The averaging system actually helps the most unbalanced player here.

Option D: All three players add another set of two, worth 54 points. Player 1 has 2,454 total, player 2 has 1,934, and player 3 has 984.

Finally, what happens when the players near the scoring cap for option B with their preferred play styles?

Player 1 has exactly 17 games for each map size.

Player 2 still only has 1 Huge game, 25 Duel and Standard games each, and 17 each of the other three map sizes.

Player 3 has 85 Duel, 1 Tiny, 1 Small, 10 Standard, 3 Large, and 2 Huge games.

Option B: Player 1 gets the full 10,200 points and cannot score any more. Player 2 loses 1,600 for the missing Huge games for 8,600 total. He can only score more points by playing Huge maps. Player 3 far exceeds the Duel cap and gets only 3,400 total points. He can score more points by playing anything but Duel.

Option C: The point cap is 10,200÷6 = 1,700 points per map size. The players have exactly the same total scores as option B. They can all increase their scores slowly by playing map sizes at the cap or quickly by playing sizes under the cap.

Option D: Player 1 still gets the full 10,200 points. Player 2 has one complete set (600), seventeen sets of five (7,480), and eight sets of two (432) for 8,512 total points. Player 3 has one complete set (600), one set of four (292), one set of three (150), seven sets of two (378), and 75 singletons (900) for 1,495 total points. This method produces about the same result as the other two for players 1 and 2 but punishes player 3 pretty severely. Players 2 and 3 can catch up rapidly by bringing their games more in balance.

 

[beezany]

Looking at the different scenarios, I see that there isn't a big difference between options C and D for mostly "balanced" players. Option D punishes "unbalanced" players more harshly, but it's much more complex to implement, so it may not be worthwhile. I definitely prefer both of them to option B after seeing the results.


– Bradd

 

[sanabas]

I actually like it as a concept. Not sure if I like it more or less than the capping at averages. It's probably more elegant, but I think it leads to a bit more gaming of the system. For instance, if you're looking at the 4 different speeds, you could play 3 games on quick/std/epic, then a fast duel/marathon game would add over 35% to the score for the first 3. Using average has the problem that it encourages playing a higher percentage of duel games, because they will probably score less, and to get the most benefit from your games you're looking to balance by total score, rather than volume. I can see a situation where you play lots of normal games, but then crank out half a dozen 5 minute games just to ensure your normal games get full points.

The downside I suspect would be implementing it, and understanding it when looking at what your future score might be. Simply counting best x games for a category is easy to understand, it'll likely be unbalanced to start, playing lots of the same game will see your score increase quickly. It'll only start to have an impact when you near some of the caps.

 

[beezany]

That's an interesting point you make about the averaging system (option C). For my comparison post, I assumed that all games were worth 100 points, but of course that's not going to be true in practice.

If easy games score less than hard games, then you'll need to play more easy games than hard games to maximize your score under option C, which is a bit counterintuitive if the goal is to reward skill and diversity. You could ameliorate this by applying difficulty modifiers after the averaging cap. That would make the scoring a bit less transparent and harder to implement, but probably still fairly simple.

Regarding the logistic curve (option D): Part of setting this up includes scaling the curve appropriately for the categories. Tempi is a weird one because there are only four choices. You can score the steps 12/50/88/100%, but really anything with the first step less than 25% is a valid S-curve. If folks like the S-curve idea, I'll put some more work into recommended scaling factors for the different categories.

 

[beezany]

I made a spreadsheet to calculate logistic curves for all of the VVV categories with 0% completion at S(–3) and 100% completion at S(+3). The numbers for Go the Distance are slightly different from above because I re-scaled the formula so that 0% and 100% would actually fall on the curve at S(±3) instead of infinity. (Map Quest and League of Nations omitted for brevity.)

Category | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8
Inferno | 0.000 | 0.053 | 0.149 | 0.302 | 0.500 | 0.698 | 0.851 | 0.947 | 1.000
Distance | 0.000 | 0.079 | 0.245 | 0.500 | 0.755 | 0.921 | 1.000
Machiavelli | 0.000 | 0.104 | 0.339 | 0.661 | 0.896 | 1.000
Tempi | 0.000 | 0.149 | 0.500 | 0.851 | 1.000

One difference between this and the averaging method is the impact of fast, low-scoring games, especially for the big tables like Map Quest and League of Nations. Because you only get full score for complete sets of maps or leaders, it may encourage people to play full-length games for their favorite settings and then fill out the rest of the tier with cheap games. In contrast, cheap games don't really help you with the averaging method (unless you play a lot of them).

Then again, I think most people use fast, cheap games to fill out the bigger VVV tables anyway. Once you do that, there isn't such a big difference between the logistic curve and the averaging method. In both cases, you can initially fill out the table with cheap games, then go back and play quality games to build up your score. As long as you have good variety in your quality games, your score should increase steadily.

 

[beezany]

Here's an example of what I'm talking about in my last post. I'll use Inferno for my example because the number of settings in the category is manageable. For the sake of discussion, I'll ignore scoring modifiers for difficulty and just assume that a “cheap” game (e.g., Quick Duel) is worth 10 points while a “quality game” (e.g., Standard) is worth 100 points. The player's scoring strategy is to first fill out the category with “cheap” games to complete VVV and then to go back and play “quality” games to maximize score. Each row has the score for the individual game, followed by the unmodified running total, the score capped with the averaging method, and the score adjusted with a logistic curve.

Difficulty | Points | Total | Avg Cap | Avg Total | Log Mod | Log Total
Settler|10| 10 |1.25| 1 |0.053| 1
Chieftain|10| 20 |2.5| 5 |0.140| 3
Warlord|10| 30 |3.75| 11 |0.302| 9
Prince|10| 40 |5| 20 |0.500| 20
King|10| 50 |6.25| 31 |0.698| 35
Emperor|10| 60 |7.5| 45 |0.851| 51
Immortal|10| 70 |8.75| 61 |0.947| 66
Deity|10| 80 |10| 80 |full| 80
Settler|100| 180 |22.5| 93 |full+0.053| 171
Chieftain|100| 280 |35| 130 |full+0.140| 263
Warlord|100| 380 |47.5| 193 |full+0.302| 359
Prince|100| 480 |60| 280 |full+0.500| 460
King|100| 580 |72.5| 393 |full+0.698| 565
Emperor|100| 680 |85| 530 |full+0.851| 671
Immortal|100| 780 |97.5| 693 |full+0.947| 776
Deity|100| 880 |110| 880 |full+full| 880
Settler|100| 980 |122.5| 893 |f+f+0.053| 971
Chieftain|100| 1,080 |135| 930 |f+f+0.140| 1,063

I like the way the logistic method keeps the modified score very close to the full total once the player has a good, balanced set of games. In this example, every “quality” game adds at least 91 points to the total score. In contrast, the averaging method punishes you harshly when your score is in balance and you try to push up the cap. Because of the way the averaging works, the last “cheap” game adds 19 points to the total, but the first “quality” game only adds 13 points. That's counterintuitive and possibly frustrating for players.

I'm not sure whether that justifies the additional complexity of the logistic method, but it's starting to look more attractive to me now.

 

[TractorBoy]

OK, so I am a simpleton. I don't quite figure any of this mathematical jargon that you guys are going on about.

To fully understand this VVV scoring structure that you guys are devising would cause utter confusion to us mere mortals. I don't even understand the mean when associated with RA's let alone this load of old gobbledegoop!!

Why not keep it simple?

 

[beezany]

That's a valid criticism, TractorBoy! While I really like the results from the curve method, it's tricky to implement and explain in detail. The averaging method is simpler, but it has some quirks too.

However, they both share a pretty simple basic principle: You get the best scores by playing a balanced mix of settings, the worst scores by playing the same settings over and over.

 

[FeiLing]

I don't bother either with all the stuff that has been written here, just hoping whatever you do kinda works out in the end.

So... can anyone actually explain to anyone else in 2 sentences what this new scoring system is?
Or will I when i go to the FAQ once it's done only get to see those tables I currently see in this thread (which make me say "oooookay... whatever.")?

 

[beezany]

So... can anyone actually explain to anyone else in 2 sentences what this new scoring system is?

The balancing system rewards players who demonstrate mastery of all the different game settings. For example, you will score full points for Go the Distance if you play all six map sizes about equally, much less if you predominately play one map size like Duel or Standard.

(The discussion and tables and examples are to compare the different options for accomplishing this. I am hoping that the math-minded folks can reach a consensus on which option is best and then figure out a good, simple way to explain it so that the FAQ isn't full of math and tables.)