MP Ratings / Challenge Matches Open Thread


#101

Most of the events that earn players enough points to make the top 250 are best game plus match play. Of those in the 2017-2018 circuit, I think only 4, the 24 hour, Pinburgh, City Champ and Cactus Jacks, were match play all the way. I know PAPA, INDISC, Pinvasion, Texas, Pittsburgh, Louisville, Pinmasters, NW, etc. were all qualifying + match. So were the larger non-circuit events like the NW Show, Zapcon [Phoenix], Free Play Florida, Houston, Allentown, etc.


#102

You’re comparing apples and oranges. It doesn’t really matter where WPPR points come from. You should be thinking in different terms. What’s important here is: Of the players MP Ratings are rating, are they playing a sufficient amount of match play games to rate them accurately.

What I’m saying is that generally speaking players are playing enough match play games to rate them accurately and that large best game tournaments generally speaking are not where players play the most of their competitive games.

Either way it doesn’t matter: MP Ratings is a way to rate players based on their match play results. It’s getting better and better at doing that job as more results are submitted into the system. Doesn’t really matter how people play in best game scenarios since the system was never meant to rate players based on that kind of play.


#103

2017 Percentages

Funny you should mention Birmingham Tilt League.
I had probably my best year of competition in 2017, winning over a quarter of all comps I entered and reaching the finals in over half.
I was interested to see how I performed compared to all of the other UK players. I thought the comps at Tilt could be skewing the percentages somewhat, so I compared the figures with and without Tilt comps being included.
There were 12 high score game comps (Just 2 games required to win) played and a single League final (based on 5 months of playing and a final).

My figures are definitely boosted by playing against a weaker field than the other larger comps, but not as much as I’d imagined they would be.


#104

We’re going to be streaming a 7 game IFPA challenge match tonight at 8 Eastern if you want to tune in and check it out: http://www.twitch.tv/buffalopinball

Follow the results here: https://matchplay.events/live/challenges/253


#105

@pinwizj: the always hardworking @kdeangelo has fed in some older tournament data (including old INDISC results). Elwin is now up to #44 while you’re still floundering in the 400 range because of a lack of results. Agreeing to that Challenge Match against @PinballProfile cost you 21 rating points. :slight_smile:


#106

The things I do for charity :slight_smile:


#107

I have issued 3 challenges, but have no reply to either of them yet. How is the challengee notified that someone wants to play them?


#108

@JimiWolf They have to go to https://matchplay.events/live/challenges and they’ll be listed. No notification is sent because I figured that people would be standing next to each other when dealing with this and the notification would just be annoying…


#109

Easiest. Win. Ever.


#110

I vote for a single notification to pop up since the challenge could be made long distance too.


#111

Would be a great thing to add as a suggestion on the feedback forums: https://matchplay.uservoice.com/


#112

Is this a chat bot now, or do you type this every time?


#113

I’m a hard worker. I do all my own typing. :slight_smile:


#114

everything looks fine to me…:wink::sweat_smile:
I didn’t even know about this metric until today, lol.


#115

What does this mean? I just played 4 challenges in the same evening with someone we were enjoying it so much. Does that mean the last 3 don’t count?

Further, exactly what is the current impact of a challenge? My understanding is that it only affects MP Ratings and nothing on the IFPA side of things, although that may change in the future.


#116

It all counts as long as both players agree to the challenge matches being played.

Correct. Like I mentioned earlier in the thread the hope is that we’re able to drive people during 2018 to shove a bunch of data through the MP sausage grinder, as it makes for a more accurate metric. From there Andreas and Brian can talk about the implementation side of things between MP and IFPA.


#117

Hey Andreas,

I have a gut feeling that these ratings are inadequate in match play situations. I just want to present some anecdotal evidence. I think this is a shortcoming of glicko instead of a bug.

Taking a look at the 24 Hour Final Battle 2018. This was a 100-player tournament that had 16 rounds total, with 3 games played per round, in 4 player groups. That’s a lot of data!

Just taking a look at the top 5 players. Here are their point totals and their rating +/- over the tournament.

Result Player Name Points Tiebreakers
1 Trent Augenstein 93 pts. 20:11
2 John Delzoppo 92 pts. 17:16
3 Tim Sexton 88 pts. 19:11
4 Bowen Kerins 88 pts. 16:15
5 Zach Sharpe 86 pts. 19:10
Result Player Name Starting MP Rating Ending MP Rating Differential
1 Trent Augenstein 1700.73 1721.33 +21
2 John Delzoppo 1733.19 1729.57 -4
3 Tim Sexton 1686.28 1684.36 -2
4 Bowen Kerins 1804.37 1744.83 -60
5 Zach Sharpe 1656.03 1664.55 +9

Overall, my gut feeling would be that for all of these players, there would be a + next to their name. The tournaments with difficult match play portions where you’re reseeded often and have to face more difficult competition seem to be more flattening than the straight head-to-head matches.


#118

I disagree. You’re using the largest tournament in the database (apart from Pinburgh) – pick a tournament that’s not an insane outlier and look at those results. :slight_smile:

Very large tournaments will give volatile results – that is the nature of glicko. This is especially true when it’s relatively few players playing many matches. The Pinburgh results are processed with each round of Pinburgh as if they were separate rating periods to minimize this effect. I haven’t done that for the 24-Hour Battle, but it’s worth considering.


#119

Nope. I played well below my rating at 24 Hour Battle this year. You and Delz played just about right. I find it hard to believe Zach played above average, but the system might not have all his data.

Ratings go up or down based on the quality of your play relative to expectation. According to MP ratings, I should have played better (I was 1800+) so I lost rating. MP still has me higher-rated than any of the other top 5 here, but it also has some strong evidence that I didn’t deserve that high rating!


#120

You may have, but also it could be a partial function to how often you play Matchplay-linked tournaments/how far back the Matchplay system logged to determine the initial standings/an anomaly caused by Supressed Player Syndrome due to IFPA integration.

@haugstrup could probably help on that one - I feel like the 80-point swing may have something to do with Glicko having low confidence in Bowen’s rating in general than just performance.