Introduction to the Pinball Statistics Power Rankings
My goal in creating the Pinball Statistics Power Rankings is to use the existing statistics available through the International Flipper Pinball Association and matchplay.events to analyze the strongest competitive pinball players’ recent performances. Every Monday, the Pinball Statistics Power Rankings will be posted here. I will also provide commentary and analysis to highlight player accomplishments at past tournaments and to highlight upcoming pinball tournaments.
Power Rankings are commonly used in sports commentary by bloggers, radio hosts, television presenters, and fans alike. There are no hard and fast rules for how power rankings must be created. Often, individuals will just subjectively rank a list of teams or players based on how they feel. Sometimes, a consensus of so-called expert rankings are used. Other times, somebody will make some spreadsheets and objectively create their power rankings by pulling in statistics and then making up a formula to combine them all together. This is that kind of power ranking system, where all the data objectively comes from sources that track actual pinball performance, but then is subjectively pumped through some spreadsheets and formulas until I like how it looks. Although I will never explicitly tell you what that formula is, I will keep it pretty consistent from week to week. As far as I know, nobody else is making any competitive pinball power rankings, which means I get to make up all the rules.
There are currently 5 statistics that I analyze in order to create the Pinball Statistics Power Rankings. They are:
IFPA World Pinball Player Ranking Points
This is the big one. This is the reason competitive pinball players wake up every single day, hop in the shower, brush their teeth, drive to their jobs, sit in an office reading emails, and then drive back home to flop down on the couch for the rest of the evening to watch television while scrolling through social media. They force themselves to suffer through this cursed daily ritual in order to earn enough money to cover the expenses required to play in pinball tournaments.
The IFPA WPPR points system is designed to be addicting. I mean that in the way mobile game advertisements brag about their games being addicting. After each of the 1st 20 tournaments a player competes in, they get to see a number next to their name drop lower and lower indicating their global rank is improving. WPPR points are a positive-only points system, where players will never lose progress by playing more. Since a player’s best 20 events count, there are no downsides to playing more frequently. In fact, doing so guarantees that a player will grow their WPPR point total, which will mostly likely vault them up the leaderboard.
There is, of course, a complex set of rules to help determine how WPPR points are distributed, which is well outside the scope of this blog post. If you really want to get down into the details, you can head over to The WPPR v5.7 rules page and read them over yourself. The important takeaway here is that since WPPRs are the main statistic used to determine qualification for various championship series, as well as bragging rights on tinder profiles and footnotes on the bottom of resumes, they are always a topic that is hotly debated on tiltforums.
Pros:
The system has been refined so much over the years with input from the player community that it has almost universal buy-in. The opt-out rate from the rankings is extraordinarily low among active competitive pinball players. The positive-only points system rarely, if ever, discourages players from playing in more events.
Cons:
Criticisms of the IFPA WPPR rankings always focus on opportunity and equity. Sure, you can get into the nitty-gritty details on how you actually know that player X from Toledo, Ohio, USA is better than player Y from Luleå, Sweden, but nobody admits it’s perfect.
Impact on the Pinball Statistics Power Rankings: Moderate
IFPA Rating Lower Bound
This is the first of the “throwaway” statistics that appear below the player bio on the IFPA player pages. The reason I call this the IFPA Rating Lower Bound, is because the value that shows up on a player’s page is the lower bound of the Glicko-based IFPA rating statistic, even though this is not explicitly stated in any documentation on the IFPA’s website.
The IFPA rating is calculated by comparing a player’s finishing result against up to 64 other players (32 higher, 32 lower) who competed in the tournament, and treating final ranking position as head-to-head wins, losses, or ties for the sake of Glicko. The IFPA website points out that these are not actual head-to-head results, and I have referred to this statistic as meaningless on forums in the past.
Pros:
This is usually the first statistic that indicates when a player has outgrown the local competition and needs to start hitting the regional or international circuit. Not that they will get any financial support from a player’s organization or sponsor to do so. That stuff doesn’t exist in pinball yet.
Cons:
This statistic doesn’t really mean anything. Let’s say I play in a 4-game match play tournament at a local bar and I come in 4th place out of 23 players. The next day, I fly to California and finish 23rd out of 196 players in The Open - IFPA World Pinball Championships. My head-to-head record submitted to the Glicko system after the bar tournament is 19-3 or .863. My head-to-head record after The Open is 32-22 or .593. Having done both of these things, I can assure you the bar event results really should not be compared as if they’re just as important as the major event results. While Glicko will factor in the strength of all the opponents that were theoretically faced in the tournament, it’s still going to trend toward overall win-loss record when calculated this way.
Impact on the Pinball Statistics Power Rankings: Lowest
IFPA EFF percentage
I believe this stands for Efficiency percentage. Although it is not documented on the main site, the calculation for this statistic seems rather simple. First, you take the total WPPR points earned by a player at all the pinball tournaments they have played in the past 3 years without factoring in any point decay for tournaments greater than 1 year old. Then, you divide that number by the total of the WPPR points that 1st place earned in each of those tournaments. Your result is your efficiency percentage.
Pros:
This stat is good to look at when paired with WPPR points because it can start to tell you how hard a player had to work for their WPPR points. Players with high EFF percentages grab WPPRs more effortlessly, while players with low EFF percentages usually need to scounge together WPPR points from many more tournaments.
Cons:
On its own, this statistic is really boring. Most people probably look at this, shrug their shoulders, and move on with their lives. Additionally, with the way the dynamic WPPR point distribution works out, a tournament like Pinburgh with 1000 players would give every player below the A Division cut line less than 10% efficiency.
Impact on the Pinball Statistics Power Rankings: Low
IFPA Power100 Win Percentage
Only the top 250 IFPA ranked players get a chance to look at how they rank in this statistic. It shows your head-to-head record over the past 3 years against every other player in the top 250. I think this statistic is fairly useful for evaluating the skill level of players who live on different continents and do not get to compete against each other very often. I also like using this statistic because it explicitly bounds my data analysis to only 250 people, instead of the tens of thousands of active competitive pinball players.
Pros:
Data is time-bounded (3 years) and is more likely to account for the more competitive pinball tournaments which feature the strongest competition. Most of the results come from larger tournaments rather than low TGP local tournaments. It also can help offset the differences in WPPR-opportunities because it’s a win percentage.
Cons:
There is a massive difference in the number of head-to-head matches compared among players even among the top 250 ranked players. If a player wants to optimize their Power 100 percentage, their incentive is to play in fewer tournaments, and especially avoid tournaments with match play or head-to-head qualification in favor of tournaments with high score qualification.
Impact on the Pinball Statistics Power Rankings: Highest
Match Play Ratings Lower Bound
I have to give a lot of credit to Andreas Haugstrup Pedersen, the creator of matchplay.events for taking on this project of creating Match Play Ratings from his massive database of game results. This is unquestionably the most complete database of actual head-to-head pinball results that has ever been created. With the high adoption of the the matchplay.events software in competitive pinball today, nearly every player has a substantial and reliable data set of results that are used to calculate their Match Play Rating.
The reason I use Lower Bound is to emphasize the time-bounded aspects of the rating statistics to create more dynamic results from week-to-week. If a player doesn’t play as often, their lower bound will decrease over time and their position in the Power Rankings will be negatively affected.
Pros:
Unlike IFPA rating, the data only includes real head-to-head results. The data is updated constantly with almost every tournament in the world.
Cons:
Glicko.
Glicko and Glicko’s cousin Elo have been written about millions of times. If you know nothing about them, I recommend you start reading about them on Wikipedia and then go from there. The fact of the matter is that as of now we live in a world of big data and online multiplayer where Glicko/Elo form the basis of nearly all online matchmaking games.
Let’s talk about how Glicko is adapted for pinball. One unique aspect of pinball when compared to other sports is its support for 4-player matches where each player achieves a score and that score is compared to the other 3 players in that match. Most sports and video games place either 2 individuals or 2 teams head-to-head in any given match. Elo, which was created for Chess, only supports 2-player head-to-head results. Glicko has the same limitation. This means that every 1 4-player pinball game is submitted to Match Play Ratings as 6 (4C2) head-to-head matches.
Example:
A tournament starts and 4 players play a 4-player game of Jungle Queen. They record the following scores.
Player 1 - 43,260
Player 2 - 68,270
Player 3- 72,340
Player 4 - 41,100
Each player, having only played 1 game, now has the following head-to-head record submitted to Match Play Ratings.
Player 3 - 3 Wins, 0 Losses
Player 2 - 2 Wins, 1 Loss
Player 1 - 1 Win, 2 Losses
Player 4 - 0 Wins, 3 Losses
The obvious issue is that each of these players played 1 game and not 3. The results are massaged in order to deal with Glicko’s limitations.
By the time they have read to this point a Glicko-head is already firing off tons of excuses. “But it will all average out over time.” “That doesn’t mean Glicko is not predictive!” “What do you want to do, implement TrueSkill and record every single player score on every game? And then you have to train the rating system on every single pinball machine that exists, all to get maybe a 5% improvement on the “accuracy” of a rating system that isn’t even implemented in the IFPA rankings or anywhere else? Andreas does so much work for the pinball community only for you to demand these increasingly tedious statistics all for your stupid pinball blog.”
To them I say: “Look buddy. Calm down, take a breath, relax. I’m not asking for anything. I don’t hate Glicko. I don’t hate Andreas or match play. I just need to acknowledge that submitting pinball results this way is a little goofy.”
