115 replies to this topic

[TrentTheWanderer]

I've been telling folks about your idea Roland, and so far the response is really good!

[Karl Streiger]

Love it - is pure simple - every good system has to be simple - over complicating things is no warranty that it will work.
The only mistake that can be made is to think there is something like a perfect system.
Perfection doesn't exist.
Even BV - had some serious issues and i don't talking about abusing the gaps. Once I said to my self - a good commander has to win on any ground with any forces.
So - I played a couple of games - in MegaMek - the only fixed variant was BV - and the numbers of Mechs: anything else was random:
random mechs - for friend and foe
random terrain.

There were several matches - where it was obvious that victory was nearly impossible - not without luck. (short range mechs on open maps, Mechs without jump jets on the river delta map vs... medium mechs with jump jets....)

And here - I did remind myself at Sun Zi: I didn't know myself, I didn't know my enemy - and I didn't know the terrain.
BV works well - when you can choose your troops - including the knowledge of the battlefield + a educated guess about the enemy you will face.

If you reflect that on MWO - you have exact the same situation:
You hardly know about your team, you don't know the map, you don't know anything about the enemy.

Every BV attemp won't change that. So even with 100% equal teams in terms of BV, tonnage or classes -> will hardly achieve exact results. So you can't say if the BV is working or not.
Thats a big problem if you ask me.

[Trip Hammer]

Ok, I may have missed this same argument and if I did I apologize as I'm at work and have limited time.

First I really like the basis for your idea. However, you will have problems that will crop up with usage in a couple of ways.

• As new Mech types as they become available, people spend a lot of time playing them to get the skill unlocks finished. This results in them being heavily used while other more useful Mechs will have their usage drop. This will only have a limited time effect but it will lead to periods where very strong builds will have artificially lowered values and the new mechs will wind up with spikes in their values. Once again it will normalize but there will be some wonkyness when they drop new mechs.
  
• As Mechs values rise people will likely be forced to use other Mechs if they are in groups so that they can get in under whatever BV limit is set. This will in the long run drop their higher values and improve the values of inferior Mechs. I have no way to know how much this would actually affect your BV's (and I don't believe you could either without being able to ACTUALLY run the system) but I believe it would affect it and in a negative way at that.

I don't have a solution to the problem that I have presented here but I though I would point it out anyway.

[Roland]

Good comments.

The Faceless, on 07 March 2014 - 09:49 AM, said:

• As new Mech types as they become available, people spend a lot of time playing them to get the skill unlocks finished. This results in them being heavily used while other more useful Mechs will have their usage drop. This will only have a limited time effect but it will lead to periods where very strong builds will have artificially lowered values and the new mechs will wind up with spikes in their values. Once again it will normalize but there will be some wonkyness when they drop new mechs.

It's true that when new mechs show up, their prices will start at 0. This, combined with their novelty, will result in their usage being very high. This will, in turn, cause their price to rise up quickly as well, which will fairly quickly equalize. Although for a period, as you say, this equilibrium will likely be higher than normal due to the novelty of the mech and the fact that people will be dropping it for purposes other than simply winning.

However, I do not believe that it will result in quite the effect you describe here. It won't really cause really strong builds to get particularly cheap, because the effect of the new mech getting dropped a lot will be distributed evenly over ALL the other variants in the entire game. Thus, all other mechs will have their prices slightly depressed for a short time after a new mech is released, but this effect should be small enough as to not really cause significant problems.

The Faceless, on 07 March 2014 - 09:49 AM, said:

• As Mechs values rise people will likely be forced to use other Mechs if they are in groups so that they can get in under whatever BV limit is set. This will in the long run drop their higher values and improve the values of inferior Mechs. I have no way to know how much this would actually affect your BV's (and I don't believe you could either without being able to ACTUALLY run the system) but I believe it would affect it and in a negative way at that.

Yes, as inferior mechs are dropped, their prices will rise. This is really the point of the system.

However, a really REALLY bad mech (say, a locust) would start out having a very low value. At that point, someone may use it at that very low price, in order to drop another high value mech like a highlander.

However, the two mechs are probably never really going to equalize in price... Because as the locust starts getting used by players, its price will rise gradually. But quickly, while its price will still be lower than the highlander, it'll soon be equal to some OTHER mech, which is worse than the highlander but BETTER than the locust. At this point, the playerbase would start dropping whatever THAT mech is, instead of the locust.. And then the locusts' price would reach an equilibrium at that low point.

In a market where only the highlander and locust existed, it would be closer to the effect you suggest.. However, in a complex market with tons of buyers selecting between such a large number of variants, really bad and really good mechs likely will never equalize in usage and thus price, because of the wide spectrum of mechs. Although the system will push their usage numbers closer together (because that's the point of the system).

Additionally, there's one other factor which will help eliminate the issue you describe, and that is the mechanism I describe where solo players get inflated BV allocations, as do small teams, etc.

This factor will result in good mechs getting dropped more often than their raw price would normally dictate.. Which should help create a constant price differentiation between mechs of different utility levels.

[Roland]

I added a section to the second post presenting a few loose ideas about how this system could potentially be applied to clans and mercenary contracts in the future, if CW develops.

I think that's pretty far off in the future, but there was some interest from some folks here about that, so I figured I'd add it in up front, at least notionally.

[shotokan5]

I have nothing but the upmost respect for all the work and thought you have done. Talk to those who really played Battletech 3025 then you will understand what I am talking about. This is not a complicated problem. Get with them and they will show how simple the answer is. Best wishes, great sig also.

[IceSerpent]

Roland, what if this system is applied to equipment instead of the mechs? Let's say, we include the following:

1. Mech speed - n BV per 1 kph (takes care of all engines)
2. Heat efficiency - n BV per 0.1 heat dissipated/s
3. Armor - n BV per armor point
4. Weapons - n BV per weapon
5. JJs - n BV per JJ in each JJ class
6. Other equipment (ECM, BAP, etc.) - n BV per item
7. Modules - n BV per module
8. Unlocks - n BV per efficiency

(upgrades like FF and ES don't need to be included as they change available space/tonnage that is used by other things, DHS/SHS don't need to be included as they are covered by heat efficiency).

Now, for all this stuff above, we set initial BV (n) to, let's say, 10 points and let your system adjust that BV based on usage. For speed, heat efficiency, armor, and JJs we can calculate the distribution within Tw and get the "weight" of a particular characteristic - i.e. if most people drop in slow mechs speed value needs to go down, if most people drop in fast mechs speed value needs to go up. For the rest we can do the same thing you are doing for mech variants, but only count one weapon of a given type per mech - if you have PPCs, it counts as +1 to frequency regardless of how many you have, as having more than one means that you're not packing other stuff in that space/weight.

Just an idea that might resolve both the issues with assigning BV to a mech variant and with having to manually calculate BV for multiple pieces of equipment.

[Deadmeat313]

I love the idea! This could really work.

Also, it could be expanded upon to act as the base for further balancing / flavour adjustments in the game. For example - when CW is implemented - Factions that own a particular Mech factory could grant their faction members a small BV discount for using that Mech. This would encourage a national flavour in CW teams.


D313

[Asmudius Heng]

Just bumping this because ever since Roland first presented the rough idea i have been a massive fan.

Just remember people that EVERY system can be broken and gamed etc including this if people really wanted to ... but I think the simplicity of usage and peoples desire to win overall make this a harder system to break or manipulate.

+1 again and again and again ....

MWO needs some sort of sense of logistics and this kind of provides that. In a universe that is suppose to have more expensive mechs to run, repair, or buy ... and a limited number of them etc this gives a kind of stab at a rough way of dealing with it.

Like Rolands Ferarri example the same can be said of extremely good variants ... they are expensive to take because of supply and demand, you may have "bought" a mech but the parts and upkeep and ability to transpot etc that cannot be shown in a game like this can be taken into accouint with this dynamic battle value.

Mechs that might cost the same cBills but perform totally differently in the hands of the community now have a place ... that cheaper one might be considered the 'workhorse' mech - not because it is actually dropped a lot - its actually cheap because people shy away from it so a pilot who is skilled enough can take it on the cheap and perform well.

The community playing is the largest sample size we have for detemining whats good via whats being actually played.

[Roland]

IceSerpent, on 08 March 2014 - 11:00 AM, said:

Roland, what if this system is applied to equipment instead of the mechs? Let's say, we include the following:

1. Mech speed - n BV per 1 kph (takes care of all engines)
2. Heat efficiency - n BV per 0.1 heat dissipated/s
3. Armor - n BV per armor point
4. Weapons - n BV per weapon
5. JJs - n BV per JJ in each JJ class
6. Other equipment (ECM, BAP, etc.) - n BV per item
7. Modules - n BV per module
8. Unlocks - n BV per efficiency

Now, for all this stuff above, we set initial BV (n) to, let's say, 10 points and let your system adjust that BV based on usage. For speed, heat efficiency, armor, and JJs we can calculate the distribution within Tw and get the "weight" of a particular characteristic - i.e. if most people drop in slow mechs speed value needs to go down, if most people drop in fast mechs speed value needs to go up. For the rest we can do the same thing you are doing for mech variants, but only count one weapon of a given type per mech - if you have PPCs, it counts as +1 to frequency regardless of how many you have, as having more than one means that you're not packing other stuff in that space/weight.

I believe that a system could potentially be developed which would include a more detailed breakdown of mechs into component parts as you describe. However, the system would be far more complex to implement.

One prime reason for such a system being more complex, is that the frequency calculation would be much harder to correlate with BV.

In the system I presented, frequency ends up being quite easily correlated with a "buy" decision on the part of a pilot. Each pilot can only choose one mech each drop, and that choice precludes all others. Thus, in some ways it can be considered both a purchase and a sale from the perspective of the system. It effectively creates a direct correlation between that action and the perceived utility of the mech.

In a case where you are looking at a specific weapon or piece of equipment, it becomes much more complex. Any given configuration can drop multiple numbers of any given weapon, as well as combinations of weapons. This means that usage of one weapon doesn't necessarily mean a choice OVER another weapon, thus it weakens the correlation between usage frequency and relative value of different weapons.

I actually have thought about exactly what you describe, and did some investigations of such a system before I moved to the simplified system I presented here. Again, a system that accounts for the things you suggest could potentially be created, but would be significantly more complex to actually implement.

However, there is hope.

My hypothesis is that all of those things are largely accounted for in the simple decision to drop one mech or another. That is, the system I've described will effectively account for all of those internal variables, intrinsically, simply by tracking mech usage.

The reason for this is because selection of a given mech chassis tends to hinge entirely upon those various capabilities. That is, you choose a mech based on what it can carry, combined with other inherent aspects of the mech's chassis (hitbox geometry, etc.).

As an example, consider the ECM module. We could track the usage of the ECM module itself, and try to establish a BV for it... But doing so is largely unnecessary, because the perceived utility of the ECM module is going to present itself in the usage rates for the mechs that can mount it. If ECM is useful (it is), then we will see higher usage rates in the mechs that can mount it (we do).

Thus, by simply tracking the usage of individual variants, we indirectly track the usage rates of various pieces of equipment.

One thing that helps make this work even better than it otherwise would, is the existence of multiple variants of the same chassis, which have differences in those capabilities. We have different variants which have the same hitboxes, but different hardpoints.. some have other slight modifications in their twist and turn rates, etc.

That fairly fine degree of difference between chassis variants enables the system I've presented to even better account for things like weapons and equipment, because it generally means that a particular variant will be selected specifically for the weapons loadout it can carry... Because in many cases, for a different weapons loadout, a different variant would actually carry it better.

It's for those reasons that I think the system I present in this thread will effectively capture all the benefits you would get from a more complex system that individually tried to account for all of those things, while dramatically simplifying the implementation. In many ways, it may actually account for them more accurately, because it doesn't require a human to manually represent the complex interrelationship between weapons statistics and usage. Instead the system will just intrinsically account for that automatically.

[IceSerpent]

Roland, on 09 March 2014 - 12:45 PM, said:

One prime reason for such a system being more complex, is that the frequency calculation would be much harder to correlate with BV.

In the system I presented, frequency ends up being quite easily correlated with a "buy" decision on the part of a pilot. Each pilot can only choose one mech each drop, and that choice precludes all others. Thus, in some ways it can be considered both a purchase and a sale from the perspective of the system. It effectively creates a direct correlation between that action and the perceived utility of the mech.

In a case where you are looking at a specific weapon or piece of equipment, it becomes much more complex. Any given configuration can drop multiple numbers of any given weapon, as well as combinations of weapons. This means that usage of one weapon doesn't necessarily mean a choice OVER another weapon, thus it weakens the correlation between usage frequency and relative value of different weapons.

I might be missing something, but I think that ability to use multiple of the same weapon can be disregarded. When I mount a weapon (or any other equipment for that matter), I make a "buy" decision in terms of spending crits and tonnage. Electing to have a second weapon of the same type also comes at the expense of not having something else. So, we can look at it from a perspective of "spending". I spend X tons / Y crits and get some guns, if all I got is PPCs (no matter how many), I've made my "buy" decision of not getting other stuff.

Quote

My hypothesis is that all of those things are largely accounted for in the simple decision to drop one mech or another. That is, the system I've described will effectively account for all of those internal variables, intrinsically, simply by tracking mech usage.

The reason for this is because selection of a given mech chassis tends to hinge entirely upon those various capabilities. That is, you choose a mech based on what it can carry, combined with other inherent aspects of the mech's chassis (hitbox geometry, etc.).

This is the part I am not quite comfortable with, because you never know why exactly a specific variant was chosen. Your hypothesis holds true in competitive matches, as we can safely assume that everybody uses the most effective meta-builds. In a PUG queue people can easily pick a "good" mech variant and end up with a "bad" build for all sorts of reasons. I.e. they might not have c-bills for necessary upgrades or proper engine, they may not have efficiencies unlocked, they may be dropping with a "fun build" / "balanced build", it could be the only mech they own, etc.
Basically, your team getting a Victor doesn't necessarily mean that that Victor is going to be of any use - you just might get one with MGs and Flamers.

[Roland]

Quote

I might be missing something, but I think that ability to use multiple of the same weapon can be disregarded. When I mount a weapon (or any other equipment for that matter), I make a "buy" decision in terms of spending crits and tonnage. Electing to have a second weapon of the same type also comes at the expense of not having something else. So, we can look at it from a perspective of "spending". I spend X tons / Y crits and get some guns, if all I got is PPCs (no matter how many), I've made my "buy" decision of not getting other stuff.

But it's a bit more complex than the mech usage, because taking a given weapon COULD come at the expense of not taking another one.. or not. Your "buy" options in that situation are extremely flexible, compared to the simple, "choose one mech" decision.

Every variant has different hardpoints, tonnage, and critical space. This means that the actual options which a pilot can make are quite vast, even for a single variant, so it's not really easy to correlate the usage of a particular weapon with much else.

But as I said, I believe you COULD make such a system.. it would just be much more complex. I did investigate the possibility and decided that it wasn't really worth the trouble.

Quote

This is the part I am not quite comfortable with, because you never know why exactly a specific variant was chosen.

Ah, but therein lies the power of the system.

You don't know why they chose that variant.. but it doesn't really matter. All that matters is that they chose it over other options which could have been just as easily taken.

That single choice encompasses all of the internal reasons for that choice. It carried the weapons they wanted to bring, it moved at the speed they wanted to move, it was durable, etc.

Different consumers will have different reasons for their purchases, and the market ultimately doesn't really care what they are. It merely consolidates all of those decisions, and its ultimate price is effectively a combination of all of their rationale.

Quote

Basically, your team getting a Victor doesn't necessarily mean that that Victor is going to be of any use - you just might get one with MGs and Flamers.

While this could be true in a specific case, it won't be in the general sense.

This is the key point to remember... The overall price for a mech isn't based upon a single pilot's decision making process. If that were true, then you midaswell just ask that guy what he thought the values were, and use them for everyone.

Instead, this system is taking into account EVERYONE'S views on how useful that mech is, for whatever their individual rationale may be. Now, if a million people are all taking victors in order to put MG's and flamers on them... so be it. The price will go up because the victor is really popular.. But at that point, we'd have to think that maybe MG's or flamers actually did get good, if a million people are suddenly dropping victors with them.

Ultimately, this system requires that you kind of "let go" of trying to micromanage it, and trust that the overall playerbase has some degree of understanding of how the game works.. or, at least, will learn from those members who do understand, and copy their patterns. This will result in good mechs going up in price, which should achieve our desired goals.

[IceSerpent]

Roland, on 10 March 2014 - 07:19 AM, said:

But it's a bit more complex than the mech usage, because taking a given weapon COULD come at the expense of not taking another one.. or not. Your "buy" options in that situation are extremely flexible, compared to the simple, "choose one mech" decision.

Taking a given weapon always comes at the expense of not taking something else (not necessarily other weapon), the only exception is when you leave unused tonnage. If I take a PPC, I don't use those 7 tons for bigger engine, more armor, more heatsinks, more ammo, JJs, other weapons, etc.

Quote

Every variant has different hardpoints, tonnage, and critical space. This means that the actual options which a pilot can make are quite vast, even for a single variant, so it's not really easy to correlate the usage of a particular weapon with much else.

Sure it is. Think about it from a "marketing" point of view. IRL you determine how popular a given product is by checking sales. In other words, the more money is being spent on cookies, the more popular those cookies are. In MWO this would translate to "the more tonnage/crits is being spent on engine, the more important mech speed is".

Quote

You don't know why they chose that variant.. but it doesn't really matter. All that matters is that they chose it over other options which could have been just as easily taken.

That single choice encompasses all of the internal reasons for that choice. It carried the weapons they wanted to bring, it moved at the speed they wanted to move, it was durable, etc.

Different consumers will have different reasons for their purchases, and the market ultimately doesn't really care what they are. It merely consolidates all of those decisions, and its ultimate price is effectively a combination of all of their rationale.

I hear you, it is a good way to determine what variant is more popular, but it doesn't help with the overall goal of BV system in general. The ultimate purpose is to have a way to compare whichever build I chose to whatever build you chose, figure out which one is better and match them accordingly. The popularity by itself doesn't make it possible - as you've said, "Different consumers will have different reasons for their purchases". Basically, in the current meta you probably wouldn't want an SRM boat on your team regardless of the actual mech, but would want a PPC+AC config instead, also regardless of the mech.

I guess my main issue is with tying the whole thing to mech variants only, as there's no direct correlation between having a potential for good build and actually using that potential.

[Roland]

Quote

Sure it is. Think about it from a "marketing" point of view. IRL you determine how popular a given product is by checking sales. In other words, the more money is being spent on cookies, the more popular those cookies are. In MWO this would translate to "the more tonnage/crits is being spent on engine, the more important mech speed is".

Certainly the principle is the same as the underlying concepts I've laid out, but it's a bit tougher to measure because you don't really have a "currency" that folks are spending in that manner. Even in your statement there, you acknowledge it's a combination of tonnage and critical space... And it's also hardpoint limitations, and various efficiency elements.

For instance, not taking a weapon doesn't ONLY impact the tonnage and critical space, and isn't necessarily a direct trade for some other equipment. In some cases, simply not taking another weapon has a benefit, specifically in heat efficiency. For instance, not taking a PPC may allow you to do other things with that tonnage and space, but simply by virtue of not having that weapon and using it, the rest of your mech's heat efficiency will change dramatically.

Like I said, I think that the interplay is quite a bit more complex than it appears at first glance, and trying to specifically model that interplay would end up actually making a worse system if not modeled correctly.

Quote

I hear you, it is a good way to determine what variant is more popular, but it doesn't help with the overall goal of BV system in general. The ultimate purpose is to have a way to compare whichever build I chose to whatever build you chose, figure out which one is better and match them accordingly. The popularity by itself doesn't make it possible - as you've said, "Different consumers will have different reasons for their purchases". Basically, in the current meta you probably wouldn't want an SRM boat on your team regardless of the actual mech, but would want a PPC+AC config instead, also regardless of the mech.

Consider how this would actually be applied though, in practice.

You have some team, and you are deciding how your team is going to allocate its BV points.

For some given mech, whatever that mech is, and whatever it is carrying.. your team is spending some portion of its BV allocation.

That means that, from YOUR perspective, that mech is worth whatever BV cost is associated. You are, presumably, trying to pick a mech that will actually give you that much bang for your buck.

Now, you could choose to make a "bad" config with that mech, but that's up to you... You most likely aren't going to do such a thing purposefully. You're doing it because you THINK that it's a good configuration. Thus, from the system's perspective, it's saying "Hey, this guy thinks this variant is good!"

One thing to bear in mind, which is difficult for a lot of folks (this isn't specifically directed at you, but at the wider set of readers), is that there isn't really a "true" value here. There's no "right' answer, other than what the overall market decides. This can sometimes be confusing to people, who imagine specific outcomes that differ from their own beliefs. In some cases, such outcomes wouldn't actually happen... in others, they would, because the personal values set by that person aren't actually widely held.

Quote

I guess my main issue is with tying the whole thing to mech variants only, as there's no direct correlation between having a potential for good build and actually using that potential.

Well, it assumes that each person is trying to achieve the best possible build. Their ideas about what constitutes that are different.

So, in some ways, you can consider that this isn't really evaluating the true utility of a mech, but rather the perception of utility that a mech variant has. That is, it's capable of carrying a loadout that the pilot THINKS is good. Thus, the mech is perceived to have utility.

And while many people may choose to run less than optimal builds, I think that from experience we see that the most powerful builds do indeed tend to show up pretty commonly compared to other variants, to the extent that prices will likely reflect their effective utility.

[IceSerpent]

Roland, on 10 March 2014 - 09:56 AM, said:

Certainly the principle is the same as the underlying concepts I've laid out, but it's a bit tougher to measure because you don't really have a "currency" that folks are spending in that manner. Even in your statement there, you acknowledge it's a combination of tonnage and critical space... And it's also hardpoint limitations, and various efficiency elements.

For instance, not taking a weapon doesn't ONLY impact the tonnage and critical space, and isn't necessarily a direct trade for some other equipment. In some cases, simply not taking another weapon has a benefit, specifically in heat efficiency. For instance, not taking a PPC may allow you to do other things with that tonnage and space, but simply by virtue of not having that weapon and using it, the rest of your mech's heat efficiency will change dramatically.

True, but I don't see how it adds complexity to the system. Not taking that PPC means that you lose BV of that PPC, but gain BV from increased heat efficiency plus BV from whatever you take instead of that PPC. Everything seems to be covered already. Tonnage and critical space impact is intermediate, it really doesn't matter. What matters is the result of your modifications after all is said and done, and you saved the new loadout. You decrease BV by removing things and increase it by adding new things. Regardless of what is being replaced by what, you would always end up with some BV addends getting smaller and others getting larger.

Quote

Consider how this would actually be applied though, in practice.

You have some team, and you are deciding how your team is going to allocate its BV points.

For some given mech, whatever that mech is, and whatever it is carrying.. your team is spending some portion of its BV allocation.

That means that, from YOUR perspective, that mech is worth whatever BV cost is associated. You are, presumably, trying to pick a mech that will actually give you that much bang for your buck.

All this works fine for a full premade in a lobby (i.e. a private match), but those matches are not a big concern, as two teams can always come to some sort of an agreement about what to take to the battle and how to value stuff.
We will run into problems with a regular queue, where MM comes into play, which coincidentally is the primary application of matching system of any kind. In there it won't be MY perspective, my team would be saddled with a lousy config on a good variant, because that other player thinks that it's a good configuration and MM doesn't know any better because BV only takes into account the mech itself. Just an example off top of my head - what would happen if PGI decides to make Victor a trial mech, assuming that current meta holds? It would have sky-high BV due to its popularity and match outcomes would heavily depend on which team is lucky enough to get meta builds and which one is unlucky enough to get trial builds, so we would end up in the same situation we got now, just with a slightly different flavor.

[Roland]

Quote

True, but I don't see how it adds complexity to the system. Not taking that PPC means that you lose BV of that PPC, but gain BV from increased heat efficiency plus BV from whatever you take instead of that PPC. Everything seems to be covered already. Tonnage and critical space impact is intermediate, it really doesn't matter. What matters is the result of your modifications after all is said and done, and you saved the new loadout. You decrease BV by removing things and increase it by adding new things. Regardless of what is being replaced by what, you would always end up with some BV addends getting smaller and others getting larger.

The problem isn't adding up the BV, the problem is correlating specific mech drops to an appropriate BV adjustment for all of the pieces of equipment. If you know the BV of everything calculating a total BV for the mech is trivial.

But consider the problem of looking at all of the usage stats, and trying to derive the relative frequency of every piece of equipment.

With mech frequency, I can easily derive a standard deviation, because the population size is simply how many times mechs dropped, since each choice is a selection of a single mech.

For equipment, the population size is... well, I don't even really know what it is. Because it's not just the weapons used, right? Since it's other qualities of the mechs too, some of which result from NOT using equipment.

So if you look at how the system I've presented works, it only needs to know the total drops, and how many of each mech dropped.. then it can make a few trivial calculations and derive a price curve. But with equipment, there's really no population size to do that math with. So you start needing to manually assign a ton of different correlations between the different factors, and that starts getting way harder.

Like I said, my gut says I could put something together to do it, but I haven't figured it out yet.. all I know is that you couldn't just count up the frequency of times pieces of equipment were used, because that wouldn't work... and if that part doesn't work, then the methodology presented here won't work. So we'd have to come up with something else.

Quote

All this works fine for a full premade in a lobby (i.e. a private match), but those matches are not a big concern, as two teams can always come to some sort of an agreement about what to take to the battle and how to value stuff.

Well, note that the primary application I present here for this system is as a mechanism for limiting mechs that can be dropped by progressively larger teams. It's more akin to a tonnage limit than a measurement tool for matching teams.

Quote

We will run into problems with a regular queue, where MM comes into play, which coincidentally is the primary application of matching system of any kind. In there it won't be MY perspective, my team would be saddled with a lousy config on a good variant, because that other player thinks that it's a good configuration and MM doesn't know any better because BV only takes into account the mech itself.

But that other guy who isn't in your group isn't really affecting your BV. He's not wasting any of YOUR Battlevalue allocation.

The only people you are trading BV with are folks you are grouping with ahead of hitting launch.

And if you are in a smaller group, then you are getting a bonus BV allocation anyway, allowing you to drop in better mechs to make up for the possibility that the other team is a 12 man premade who is well organized and is focusing much more on dropping optimized mechs.

Quote

Just an example off top of my head - what would happen if PGI decides to make Victor a trial mech, assuming that current meta holds? It would have sky-high BV due to its popularity and match outcomes would heavily depend on which team is lucky enough to get meta builds and which one is unlucky enough to get trial builds, so we would end up in the same situation we got now, just with a slightly different flavor.

Well, in that case, what you would likely see (depending on the number of solo players) the Victor's price increase when lots of solo players took it.

So in an actual match, an organized team would have to make significant sacrifices to drop victors... Because they would be paying heavilly for it, and they wouldn't have many bonus BV points allocated per pilot. (ie, your team wouldn't likely get a lot of meta-build victors, or if you did, they'd likely be played by solo players)

The solo new players you are worried about, dropping in trial mechs, would be able to drop Victors easilly, but as you say, they'd be trial mechs and generally pretty ineffective.

So basically, with the victor showing up in trial mechs, the effect is that Victors would probably go out of circulation for organized groups for a while.

OR, as I suggested earlier as a possibility, you could just not count trial mechs in the market. Dropping in them would have no impact on prices.

[Deadmeat313]

I think Roland's idea is pretty robust. The danger when adding granularity down to equipment level is that some equipment is very common across all chassis.

If every time a Mech is loaded that has Medium Lasers equipped the BV value of that component is increased, they will quickly go through the roof. An artificial balance would need to be implemented - and that could cause problems down the line.


D313

[Cimarb]

Deadmeat313, on 10 March 2014 - 11:53 PM, said:

I think Roland's idea is pretty robust. The danger when adding granularity down to equipment level is that some equipment is very common across all chassis.

If every time a Mech is loaded that has Medium Lasers equipped the BV value of that component is increased, they will quickly go through the roof. An artificial balance would need to be implemented - and that could cause problems down the line.


D313

If every time a game is loaded people are in DDCs, they will quickly go through the roof - it's the same thing as medium lasers. If everyone is equipping them, they are obviously extremely good because they are popular, correct?

Based on Roland's popularity model (I know you don't like the phrase, but that is what it is), the more something is used, the better it is and the higher it's cost, so the medium laser should have an extremely high cost because it is used on darn near every mech on the field.

Edited by Cimarb, 11 March 2014 - 06:39 AM.

[IceSerpent]

Roland, on 10 March 2014 - 07:39 PM, said:

The problem isn't adding up the BV, the problem is correlating specific mech drops to an appropriate BV adjustment for all of the pieces of equipment. If you know the BV of everything calculating a total BV for the mech is trivial.

But consider the problem of looking at all of the usage stats, and trying to derive the relative frequency of every piece of equipment.

With mech frequency, I can easily derive a standard deviation, because the population size is simply how many times mechs dropped, since each choice is a selection of a single mech.

For equipment, the population size is... well, I don't even really know what it is. Because it's not just the weapons used, right? Since it's other qualities of the mechs too, some of which result from NOT using equipment.

I think it's not quite that bad - there are two kinds of equipment:

1. discreet (has PPC / doesn't have PPC). In this case set is {0,1}, so you can get BDA and Sigma from there, just like you do for the mech variants.

2. Continuous (speed, armor, etc.). In this case BDA and Sigma can be calculated as:





where definite integrals have range from min (speed, armor, etc.) to max (speed, armor, etc.)

p(x) is probability density, for normal distribution it is



Once you get BDA and Sigma for each piece of equipment, you can just plug those into the rest of your equations.

Quote

Well, note that the primary application I present here for this system is as a mechanism for limiting mechs that can be dropped by progressively larger teams. It's more akin to a tonnage limit than a measurement tool for matching teams.

I understand that, but this considerably limits the usefulness of the system - premades can always agree on "house rules" as a workaround, PUG queue doesn't have this luxury and has to rely on MM for doing the "balancing".

Quote

But that other guy who isn't in your group isn't really affecting your BV. He's not wasting any of YOUR Battlevalue allocation.

The only people you are trading BV with are folks you are grouping with ahead of hitting launch.

That other guy is affecting the match outcome, and there's no control mechanism to affect that. I can't ask him to drop in a certain mech because he is not in the group, and MM can't ensure a proper match for him on the other team because MM has no idea that not all Victors are created equal, so to say.

Again, I completely agree that in organized/competitive drops we can safely assume that people run the most effective builds, so the system will work. It's not a huge benefit though, because players can "police themselves" in those drops (we've done it before in NBT-HC, and I think some player-run leagues in MWO are doing it now). My concern is really about the PUG queue, where MM has to do it automatically.

Edited by IceSerpent, 11 March 2014 - 07:41 AM.

[Roland]

Cimarb, on 11 March 2014 - 06:39 AM, said:

If every time a game is loaded people are in DDCs, they will quickly go through the roof - it's the same thing as medium lasers. If everyone is equipping them, they are obviously extremely good because they are popular, correct?

Based on Roland's popularity model (I know you don't like the phrase, but that is what it is), the more something is used, the better it is and the higher it's cost, so the medium laser should have an extremely high cost because it is used on darn near every mech on the field.

That's why the model doesn't really apply well to individual pieces of equipment, because the cost can't really be correlated simply to frequency.

However, with mechs, it works because each pilot is consistently making a single selection of which mech to drive for whatever he wants to do.

Edited by Roland, 11 March 2014 - 09:16 AM.