0
A New Approach to ELO
Before I begin let me preface the following with this. All numbers are place holders only and can be adjusted as needed. All equations if any are very much subject to change. I am not an algebra expert and I am sure I will miss something
For those that wish to know more about the founding of ELO here is a link http://en.wikipedia....o_rating_system . Despite what people say about ELO it can be used in MWO and is a great way to start. The great thing about ELO is you can cater it to your individual needs. http://en.wikipedia....iki/BattleForge Battleforge, A game that is near and dear to my heart (R.I.P) used the ELO system to great affect for their PvP. None of the PvP based players had any real issues with it, even new players were treated correctly. This post is going to be an attempt at putting an idea on how we can change the way ELO is not only factored but considered in MWO matchmaking.
Currently ELO in MWO is based off of your win loss record exclusively. This is how ELO was originally intended for chess. The problem with this is chess is a two player game. Currently MWO is a 24 player game. Chess you start off with even pieces and board. MWO every mech is different and every map has its advantages and disadvantages. To use strictly your win loss record to determine your ELO in MWO I believe is a major flaw, and if corrected should significantly improve game performance within a certain set of boundaries
Currently teams whether they be in the solo que or the group que are based off of the average ELO of the groups. This aspect of the MM does not need to change. The only change that is needed is to recalculate how the actual ELO number comes about. The great thing about recalculating is the formula is already in the game to do so. Match Score
The end of match score is based on many things. All of which are accomplishments that the pilot did during the match. Those items include, damage, kills, kill assist, spotting, ECM, Counter ECM. There may be more but I do not have access to the game client right now. Currently that scale goes from 0 to 150 is the highest that I have seen. The better you do, ie more you participate, the higher your match score win or lose. That is the key, your match score is not determined by whether or not you won or loss, but how you did in the match. We have all carried in matches, it happens, however no matter how hard we carry we still lose, and as such our ELO suffers through no fault of our own. I feel that the match score is a much better tool to measure a player’s skill than simply a win or a loss.
You then take the average “new” ELO based off of “Player Accomplishmet” (Covered Later) for each chassis. Not weight class but each chassis. The key here is it being applied to different mechs even the same weight class. A Jenner F handles much differently than a Jenner D or an Oxide. An Atlas handles much differently than a Victor. So the “new” ELO would be based off of the players performance per mech. This will help those of us who play Jenner’s all day long and then hop into a Stalker they rarely play. There will be a built in fixture to prevent exploiting this feature (explained later).
So with all that said here is a sample formula I have been kicking around for a while to determine a “new” ELO system for all players. I understand that the Match Score formula is already in the game, this is merely an example of who it could really be used to determine your ELO.
KL = Kill, KA = Kill Assist, DM, Damage, ECM = ECM/Counter ECM, TG = Tag, NR = Narced, DR = Duration (How Long Survived) WN = Win, LL = Loss
So (KL + KA + DM + ECM + TG + NR + DR + WN - LL) = Match ELO modifier for that chassis, well that will really inflate the numbers just on damage alone, so other modifiers need to be put into place. The damage potential or an assault is much higher than that of a light pilot. The most kills and kill assists you can get in a match is 12, so we can leave that number alone. Damage can be very high and the potential damage between a 100 ton mech and a 20 ton mech can be substantial, so a factor needs to be put in to account for this. Duration of the match survived I think is a good one as well. Did the player suicide early or did he hold out for the long run and contribute. However there needs to be a factor here as well.
So the factor for damage, we have all seen Assault mechs get 1,000+ points of damage. Of course we have all seen light pilots get 1,000+ points of damage. So how do you balance that? Well damage potential is what we should be looking at. Potentially an assault mech should always outgun a light mech. So you we have to take that into consideration. We all groan when we see direwolf only do 100 points of damage, but nothing is said about the locust that only did 50 points of damage. So based on the class we modify the damage by their weight. So DM becomes (DM/Modifier). See the Bottom for a breakdown of factors. With an average damage of 500 an atlas would have 125 added to his “new” ELO rating and a jenner would have 143 added. The atlas is expected to do heavy damage to where the jenner is not. Scripting can be called into play to determine the weight class. The calculated variable calls into play the less weight and therefore less weapons a mech can bring to bear. This also allows for those excellent matches to be rewarded.
Now factors for damage would be listed as such
20 tons 3.36, 25 tons 3.40, 30 tons 3.44, 35 tons 3.48
40 tons 3.52, 45 tons 3.56, 50 tons 3.60, 55 tons 3.64
60 tons 3.68, 65 tons 3.72, 70 tons 3.76, 75 tons 3.80
80 tons 3.84, 85 tons 3.88, 90 tons 3.92, 95 tons 3.96, 100 tons 4.0
Duration needs to be taken into account as well. If you are in a light which spawns close to the enemy then rush to their spawn point drop an Arty and hit them all then suicide you could exploit the system. So the longer you are in the match the higher this should boost your “new” match ELO score.
Tag and Narc bonuses should be weighted towards lighter mech chassis vs heavier ones. Its not generally the job of the heavy and assault mechs to tag and narc. However they do and they should be rewarded for that. The duration of tag and narc can both be tracked. So you calculate the number of seconds each was done on a target and divide by the modifier. For the ECM and Counter ECM, for every mech you counter 5 will be added to your score. A win will grant you 10 to your score and a loss will subtract 10
Examples
A jenner weighing in at a whopping 35 tons, does 565 points of damage, with 2 kills and 5 kill assists, who won the match and lasted 7:35
A raven weighing in at a whopping 35 tons does 400 points of damage, with 1 kill and 6 assists, with 240 seconds of tagging, and 210 seconds of narcing, who won the match and lasted 7:35
So applying (KL + KA + DM + ECM + TG + NR + DR + WN - LL) = Match ELO modifier for that chassis.
Jenner D (2 + 5 + (565/3.48) + 0 + 0 + 0 + 8 (Rounded Up) + 10) = 187
Raven 3L (1 + 6 +(400/3.48) + 15 + (240/3.48) + (210/3.48) + 8 (Rounded Up) + 10) = 284
While the Jenner had more damage, more kills, the raven also contributed tag, narc and ECM counter ECM to the equation, making his participation in the battle a little more than the jenner and he was rewarded for it. Now the key to this formula is keep it within the same chassis. Meaning a separate ELO for the Jenner F, The Jenner D, The Jenner K, The Oxide, so 4 ELO ratings which are separate since they are separate mechs and are played differently. Very quickly your ELO score is based off of your ability to perform while in combat and not based off of your teams’ abilities.
Well what about Clans, with their ability to swap Omni pods it can be difficult to track, so my advice would be just to use the base mech ie, the center torso :>
Well what about new players. New players should always always start at the bottom of the ELO scale and work their way up. Eventually they will fit where their skills will let them fix.
There are literally a hundred different ways to do this other than basing it off your win loss record. To me that is just a slap in the face for any player. This is merely an example of how it could be done using a system that is already in the game and I think is pretty damn fair in judging your performance based on an individual basis.
To find your ELO score all that needs to be done is average your “new” match score and viola. As I write this I have thought of a few flaws in the system already that I did not address. But this is merely an easy way to get rid of win/loss basis and make it more based on your performance in a match.
Once again all numbers are arbitrary and merely placeholders. However the system is already in play for the game, why not make it do double duty?
Edited by Darian DelFord, 09 October 2014 - 10:57 AM.
