38 replies to this topic

[Magna Canus]

A lot of complaints come in over various forum posts from experienced players centered around the performance of new players. Most agree that better tutorials, more battle experience, etc. are necessary to bring them into the ELO mid field in which new players start now.

The current assumption is, as I have interpreted it, that by starting in the mid range individual player ELO would quickly balance out to their "true skill".

The theory is sound in theory alone, but in daily practice it does more harm than good;

1) New players get placed together with experienced players and get decimated time and time again, making their experience less enjoyable and therefore the chance of player retention much lower.
2) Many experienced players resent being "saddled" with new players that have yet to master the basics of game play such as targeting, movement, noticing when they are being shot, differentiating between friend and foe, understanding the mini-map, etc. the list goes on and on. This creates tension in game that can often become very abusive and verbal in its expression and again reduces the quality of a players experience and the potential of player retention.

Therefore I would like to suggest that the "starting ELO" of new players be set to 0 (Zero). This would ensure that new players were only included in higher level matches as a absolute last resort by the MM, reducing the number of frustrating events and the level of frustration on all sides. From there a new player can "earn" the right to claim a higher ELO and be placed into matches more on par with their skill set.

Like in professional sports, chess tournaments, or any other activity that would use an ELO system, one should be required to display their skills and abilites in the minor leagues before moving on to the major league.

As far as I understand it this should be a very easy and minor change to make, the simple changing of a single number that would, in my view, bring with it huge benefits to the community and the game in general.

Edited by Magnakanus, 14 March 2014 - 03:59 AM.

[NextGame]

it wouldnt hurt

[Charons Little Helper]

Elo doesn't work like that.

People in chess don't start with an Elo of 0 either. (I played ranked tournaments for some time as a kid - though my rating never got all that high.)

If they did start new players at 0 - it would cause a massive Elo deflation - and it would be almost impossible for new players to work their way up to a decent Elo due to only playing other people around 0.

If you want to invent your own whole rating system from the ground up where people start at 0 go ahead. But it wouldn't be Elo.

[Bilbo]

Besides allowing vets to troll the new kids on an alt account, new players are better off starting in a position in which they have some chance of actually learning something from the people they are playing with/against.

[Magna Canus]

Charons Little Helper, on 14 March 2014 - 04:48 AM, said:

Elo doesn't work like that.

People in chess don't start with an Elo of 0 either. (I played ranked tournaments for some time as a kid - though my rating never got all that high.)

If they did start new players at 0 - it would cause a massive Elo deflation - and it would be almost impossible for new players to work their way up to a decent Elo due to only playing other people around 0.

If you want to invent your own whole rating system from the ground up where people start at 0 go ahead. But it wouldn't be Elo.

Hi and thank you for that information. I may need to reword the post, though the core idea remains the same; start at the bottom and work your way up. Whether the lable is ELO or clown-points is irrelevant. In order to "calculate a rating" there has to be some form of information to calculate from. As a new player you have no performance information avaiable to be calculated. In fact, due to the steep learning curve, the advantages that different levels of efficiencies bring, as well as having C-bills available to purchase the equipment you need, each new player has a number of inherent factors that impare their performance compared to a similarly skilled player with those same advantages.

Yes, a great player with poor equiment/etc. will still perform well, but with good equipment/etc. they can perform even better.

So if you will, the 0 represents as well an inherent "equiment/efficiency/C-Bill" modifier.

With an ELO of 0-x a win/loss or whatever was used at the time as an ELO influencing factor could be set in increments until one reached the "average" mark.

[Edustaja]

Yeah, average new player is not the same as an average player.

[NextGame]

Right now, the new player starts at "average"

Theres a few things that this results in.

1) Brand new players will generally not contribute effectively to matches
2) New players will get carried to some extent and will not normalise into their skill bracket for a while
3) New players won't necessarily learn as quickly if they start right at the bottom, as they will be in with bads.
4) Hypothesis: new players starting as "average" will lose most of their first few games badly and this wont create the best first impression of the game.
5) New players starting as "average" are generally going to be using inferior mechs both in terms of equipment and also in skill unlocks.

Other than newer players learning slower I dont really see much drawback to starting them at the bottom, or closer to the bottom, of the pile, as even if they are decent, they will need time to get a worthwhile mech in order, as well as get to grips with the basics of the game (not that its that hard, but nevertheless)

Edited by NextGame, 14 March 2014 - 06:26 AM.

[Charons Little Helper]

NextGame, on 14 March 2014 - 06:20 AM, said:

Other than newer players learning slower I dont really see much drawback to starting them at the bottom, or closer to the bottom, of the pile, as even if they are decent, they will need time to get a worthwhile mech in order, as well as get to grips with the basics of the game (not that its that hard, but nevertheless)

In the long term - this would cause Elo deflation - making wherever new players start become the eventual average.

From what I understand - new players are given a 200 point nerf to their Elo while on cadet bonus. If you want to argue that it should be increased - fine. But whereever new players start will eventually become the average.

The only systems I've ever seen which can start new players at 0 or whatever - are systems in which you can only earn points - never lose them. This isn't really a good system to figure out skill - just who plays the most.

[NextGame]

Charons Little Helper, on 14 March 2014 - 06:31 AM, said:

In the long term - this would cause Elo deflation - making wherever new players start become the eventual average.

From what I understand - new players are given a 200 point nerf to their Elo while on cadet bonus. If you want to argue that it should be increased - fine. But whereever new players start will eventually become the average.

The only systems I've ever seen which can start new players at 0 or whatever - are systems in which you can only earn points - never lose them. This isn't really a good system to figure out skill - just who plays the most.

If there is indeed a nerf then it might be in the ballpark of where it needs to be, with some tweaking, but PGI don't exactly furnish us with the finer detail.

[Magna Canus]

Bilbo, on 14 March 2014 - 06:03 AM, said:

Besides allowing vets to troll the new kids on an alt account, new players are better off starting in a position in which they have some chance of actually learning something from the people they are playing with/against.

Vets trolling new players: Yes, for those mean spirited persons so inclined creating a new account to troll ELO 0 players would be an issue. Depending on how cynically you view the moral fiber of humanity this becomes more or less of a real problem. The use of a cap in the form of number of games played, ELO score reached, etc. should help to make these "joys" short lived.

Leaning from your team/enemy; Learning is usually done by either study, observation, training/teaching, trial and error, or any combination thereof. For those that learn quickly thought the use of forum tips, tutorials, reaching out to units for training, observation, etc. the learning curve is less steep than for those that learn through trial and error. My current impression is that the majority of new players are learning the game through trial and error which is more effective when done at a slower pace with others that are similarly unsure of what they are doing. Pitting a new player vs. a vet and expecting that new player to learn something in the middle of getting themselves decimated is less than realistic. Yes, there are exceptional individuals with the motivation and self-confidence that can, but my assumption is that the average person casually checking the game out is not going to be one of these.

As I stated in the OP we all know that tutorials are lacking in the moment and I do not see many community offers to new players to train them that do not involve directly joining a group.

[Bilbo]

Magnakanus, on 14 March 2014 - 06:38 AM, said:

Vets trolling new players: Yes, for those mean spirited persons so inclined creating a new account to troll ELO 0 players would be an issue. Depending on how cynically you view the moral fiber of humanity this becomes more or less of a real problem. The use of a cap in the form of number of games played, ELO score reached, etc. should help to make these "joys" short lived.

Leaning from your team/enemy; Learning is usually done by either study, observation, training/teaching, trial and error, or any combination thereof. For those that learn quickly thought the use of forum tips, tutorials, reaching out to units for training, observation, etc. the learning curve is less steep than for those that learn through trial and error. My current impression is that the majority of new players are learning the game through trial and error which is more effective when done at a slower pace with others that are similarly unsure of what they are doing. Pitting a new player vs. a vet and expecting that new player to learn something in the middle of getting themselves decimated is less than realistic. Yes, there are exceptional individuals with the motivation and self-confidence that can, but my assumption is that the average person casually checking the game out is not going to be one of these.

As I stated in the OP we all know that tutorials are lacking in the moment and I do not see many community offers to new players to train them that do not involve directly joining a group.

http://mwomercs.com/...then-look-here/
http://mwomercs.com/...e-chat-servers/
^ Right there in the new player's section.
I don't know about you, but if were just casually checking things out, the last thing I would want is to be dropped in game with 23 other guys whose only response to any questions I might have is: I dunno.

[Magna Canus]

Charons Little Helper, on 14 March 2014 - 06:31 AM, said:

In the long term - this would cause Elo deflation - making wherever new players start become the eventual average. From what I understand - new players are given a 200 point nerf to their Elo while on cadet bonus. If you want to argue that it should be increased - fine. But whereever new players start will eventually become the average. The only systems I've ever seen which can start new players at 0 or whatever - are systems in which you can only earn points - never lose them. This isn't really a good system to figure out skill - just who plays the most.

I have to admit that I don't exactly understand this concept, especially in light of the new "ELO buckets" to be implimented on 29 April. Tier 1 ELO bucket actually goes from 0 to 1000.

An average is the summ of all scores divided by the number of players, so in order for the new player starting number to be the "new average" we would need a steady flow of new players on a regular basis to counter the higher scores of experienced players as well as the higher scores of new players whose scores have increased.

An increase in the "nerf" of 200 would suffice just as well as starting them at 0 as long as the nerf were significant enough (no particular number in mind in the moment).

As far as the viability of starting at 0 combined with being able to loose points, it should work perfectly fine. Starting at 0 is just that, you have nothing to loose and can only go up. If, after 10.000 battles you have not managed to improve beyond the cadet level then that is where you should remain.

I agree that the current ELO factors (Win/Loss) are lacking for an accurate portrayal of individual skill, but it is what we have until it is improved/expanded.

[Bhael Fire]

They need to ditch Elo altogether; it's not appropriate for a game like MWO.

A customized version of BattleValue would be much better. That is, a version that takes the player's mech and their individual skill with that mech into account.

Player skill needs to be rated on more than just their win/loss record.

[Bilbo]

Magnakanus, on 14 March 2014 - 06:51 AM, said:

I have to admit that I don't exactly understand this concept, especially in light of the new "ELO buckets" to be implimented on 29 April. Tier 1 ELO bucket actually goes from 0 to 1000.

An average is the summ of all scores divided by the number of players, so in order for the new player starting number to be the "new average" we would need a steady flow of new players on a regular basis to counter the higher scores of experienced players as well as the higher scores of new players whose scores have increased.

An increase in the "nerf" of 200 would suffice just as well as starting them at 0 as long as the nerf were significant enough (no particular number in mind in the moment).

As far as the viability of starting at 0 combined with being able to loose points, it should work perfectly fine. Starting at 0 is just that, you have nothing to loose and can only go up. If, after 10.000 battles you have not managed to improve beyond the cadet level then that is where you should remain.

I agree that the current ELO factors (Win/Loss) are lacking for an accurate portrayal of individual skill, but it is what we have until it is improved/expanded.

The new players are started at just below average. They have their actual ELO scores seeded through a certain period( I believe it is through the cadet bonus). Once that Period has ended they have an ELO score based on play against "average" players, which in my mind is more accurate than a score over the same period against players of unknown skill level.

[Charons Little Helper]

Magnakanus, on 14 March 2014 - 06:51 AM, said:

I agree that the current ELO factors (Win/Loss) are lacking for an accurate portrayal of individual skill, but it is what we have until it is improved/expanded.

Elo works fine in team games - it just takes longer for people to get to their equilibrium.

Sometimes you end up on a good team. Sometimes a bad team.

In the law of large numbers - your teammates / opponents will average out - the only constant is you.

Though I wouldn't be against something for groups - giving them a boost to their score for matchmaking purposes while they're grouped. (League of Legends does something similar)

Say your Elo is 1400, and your buddy's Elo is 1600. You average 1500. But since you're a 2 person group - they give you a boost - and the MM searches as if you're at 1550. (as a ballpark) Larger groups getting a bigger boost.

[Magna Canus]

Bilbo, on 14 March 2014 - 06:44 AM, said:

http://mwomercs.com/...then-look-here/
http://mwomercs.com/...e-chat-servers/
^ Right there in the new player's section.
I don't know about you, but if were just casually checking things out, the last thing I would want is to be dropped in game with 23 other guys whose only response to any questions I might have is: I dunno.

Hi, that is great info that you provided, thanks. You and I make use of the forums often and I am sure that other motivated persons are going to use them as well. The question is; how much of the total population is actually using the forums? Many people believe, myself included, that the number of people actively using the forums are in the minority, or at the very least the number of new players trying out the game who make immediate use of the forums (or maybe effectie use?) is relatively low. I base that belief on the number of complaints and personal observations of people lacking basic skills and knowledge offered here in the forums.

I am not sure if asking questions during a match is really the most effective approach, though I seen a few here and there on occasion. Normally when you are in battle you have other things to worry about than asking about how to adjust your weapon groups, etc. just like the other players. In general we have a good community that answers these questions promptly which is a big benefit, but I think most agree that a combat drop does not make the best class room when learning the basics, especially when the "teacher" is being shot at.

[Magna Canus]

Bilbo, on 14 March 2014 - 07:03 AM, said:

The new players are started at just below average. They have their actual ELO scores seeded through a certain period( I believe it is through the cadet bonus). Once that Period has ended they have an ELO score based on play against "average" players, which in my mind is more accurate than a score over the same period against players of unknown skill level.

First of all, thanks Bilbo for all the posts and participation. =)

The ground basis of my OP is one base on skill levels within specific skill ranges and does not assume that a beginner and a veteran can realistically be put on par in relation to skill, experience, and resources. Perhaps "raw talent" is on par, but under the current ELO system based off of Win/Loss experience and resources play a significant role in what constitutes your ELO. It makes a difference when you have a new mech with no efficiencies, or if you could not affort the larger reactor, the double heat sinks, etc. Founded on that, the comparison of "skill" between a beginner (new player) and a veteran (average player) is skewed.

New players need a "sand box" to learn the ropes in. Those that are "average" will leave the sand box at some time and move on. When a beginner drops they should only need to face comparable players and fairly, an average player should be able to expect a certain minimum from their team, e.g. those that have left the sand box behind them.

[Roadkill]

Suggestions like this come from a misunderstanding of how Elo works.

It isn't possible for new players to start at any point other than the middle in an Elo system. Elo is a zero sum system*, so if you move the entry point for new players all that does is move the "average" and all rankings then migrate to that new average. Doing so accomplishes nothing.

If you want new players to start low and work their way up you'll have to invent an all-new system.

But what could be done is to segregate new players (say, less than 50 matches) from experience players. Use the same ranking system for both groups, but force new players to play with each other until they have some experience and then move them into the pool with the sharks.

*not technically true in most implementations due to limits placed at either extreme, but accurate enough for our purposes.

[Bilbo]

Magnakanus, on 14 March 2014 - 08:02 AM, said:

First of all, thanks Bilbo for all the posts and participation. =)

The ground basis of my OP is one base on skill levels within specific skill ranges and does not assume that a beginner and a veteran can realistically be put on par in relation to skill, experience, and resources. Perhaps "raw talent" is on par, but under the current ELO system based off of Win/Loss experience and resources play a significant role in what constitutes your ELO. It makes a difference when you have a new mech with no efficiencies, or if you could not affort the larger reactor, the double heat sinks, etc. Founded on that, the comparison of "skill" between a beginner (new player) and a veteran (average player) is skewed.

New players need a "sand box" to learn the ropes in. Those that are "average" will leave the sand box at some time and move on. When a beginner drops they should only need to face comparable players and fairly, an average player should be able to expect a certain minimum from their team, e.g. those that have left the sand box behind them.

In MWO's case the sandbox is the training grounds. Whether anyone chooses to use them for that purpose is another thing all together. My point is, the best way to judge a players skill level is against whatever the average skill level is for the game. They won't be in that range long enough to get discouraged if they are below average nor bored if they are above. If a new guy had talent for the game I can't imagine the tedium of climbing up from zero to the middle tier, much less the highest tier.

[Trauglodyte]

Charons Little Helper, on 14 March 2014 - 06:31 AM, said:

In the long term - this would cause Elo deflation - making wherever new players start become the eventual average.

From what I understand - new players are given a 200 point nerf to their Elo while on cadet bonus. If you want to argue that it should be increased - fine. But whereever new players start will eventually become the average.

The only systems I've ever seen which can start new players at 0 or whatever - are systems in which you can only earn points - never lose them. This isn't really a good system to figure out skill - just who plays the most.

Well, what you could do is create a ladder system. Blizzard uses that now with their damned Hearthstone game. You start out with a rank of 25 and work your way up to rank 1, with varying levels once you hit 1. You don't lose any rating points until you get past lvl 20.

At any rate, Champion mech drivers and new players need to have their ELO set to a much lower level than that of everyone else. All the current system does is penalize those players and penalyze the players that are stuck with them. It ends up being horrible for everyone and it doesn't do anything to promote a safe learning experience for the people that are just coming into the game. MWO has enough hidden information and a massive learning curve. Adding to the frustration of a new player hurts more than it helps.