Amazing, a thread somewhere else that would probably fit into this forum. It was discussing how the table would've looked if there had been no overtime in MLS. I don't know how accurate any of the numbers are and finding if someone has already done a more detailed analysis would be interesting, but here's the thread. https://www.bigsoccer.com/forum/showthread.php?s=&threadid=99512
US Pro Soccer Results, Tables & "Shots at Goal" by Chris Allen Allen's page has updates through the 2001 MLS season. For each season you can find regular season results with goal scorers, playoff results, and tables. For seasons prior to 2000, the tables page includes a "Hypothetical FIFA-standard table WITH draws and WITHOUT conferences" table. Obviously this doesn't address the point about how MLS would look without OT, but it does give a look at how MLS might have looked without shootouts. I used this page a lot when, prior to this forum, we were discussing the applicability - or need for refinement - of Bill James' Pythagorean winning percentage.
That's phenominal. I was just last night trying to re-adjust my files so that I would get rid of SOW.
Does anyone have any data on how many goals were scored in OT? My impression is that rate was somewhat higher than during regulation time, but I don't know. Assuming there is a benefit to eliminating draws in league games, what higher frequency of goals would have been necessary to make the extra 10 minutes worth it? A rough calculation (ignoring the minutes of OTs where goals were scored) produces a figure of a goal every 32 minutes last year in MLS. That number applied to the ten minute overtime suggests that roughly 1 in 3 tied games would produce a golden goal. Finally, I see that there were 41 draws in 150 games last season, even with the overtime. That seems like a lot to me.
Even though it hasn't been updated in a while, that's a good site for the Resources Thread so I posted it there. I'm checking the OT goals stats right now.
Here's what I got, with a quick and careless calculation: Regulation goals:421 Regulation minutes: 13500 OT goals: 12 OT minutes: 483 Regulation goals/90: 2.81 OT goals/90: 2.24 I'm not exactly sure what NER_MCFC's other question is asking.
Your numbers look right ChrisE. I just went to MLSnet.com and counted 52 OT's and 12 goals. I think he's asking how many more goals needed to be scored in order to have both the regulation goal rate and the OT goal rate to be equal.
Interesting. Very intersting. On the face of it, that suggests that teams are less likely to score in OT. With the golden goal rule you'd expect the scoring rate for OT to be higher than regulation because it artifically restricts the number of OT minutes (the denominator). [Then again, it also restricts the numerator, too, but I doubt that has as much impact.] OTOH: you could argue that the regulation goals/90 is inflated by blow-outs and that, by definition, only closely matched teams go into OT. I would further argue that that higher scoring games are less likely to end as ties (think about volatility). Likewise, if you have two teams that play to 0:0 after 90 minutes it is unsurprising that they also go 0:0 in OT. One way to come up with a more precise answer would be to look only at games that ended in ties: determine the scoring rate for those games and calculate the likelihood of a goal in the OT time period. Then compare that to what percentage of games had goals in OT.
No, it doesn't give us a look at how MLS might've looked. It overlooks the very basic fact that teams react to the rules of the competition in which they compete. Simply ignoring shootouts or overtimes, listing teams in a "single table" is the worst type of statistics, and quite akin to numerology. And "Saint" Bill James would cry at these crude attempts.
So, Andy, if we're so terribly wrong, tell us what we should do about the shootout. What is the significant difference between how teams play when there's going to be a shootout and when the game is just going to end at 90 minutes? Visiting teams making a shell to protect their tie (or lead)? Anybody trailing throwing in pretty much everything they've got in pursuit of a chance for points? Home team pushing forward to break a tie? I don't think we ought to eliminate overtime results. But I think, much like pk's, shootouts are pretty much random, and taking them out gives you a significantly more accurate picture of how good a team was.
ChrisE, you make a good point assuming that shootouts results were indeed random outcomes. Without having looked at any of the stats my intuition is that the home team was more likely to win a shootout and there was a big difference between the very best and the very worst shootout teams (due to better goalies and/or better offensive players). Complicating matters further, even if there was no difference in outcomes as long as there was any perceived difference then coaches and teams would play adjust their strategy accordingly.
I don't have any evidence about the home/away thing, but here are coaches' shootout records. Since coaches had (mostly) the same offensive players and keepers, this table should show those effects, too. http://www.mlsnet.com/news/archive/coaches.html Coaches who went through at least 10 regular season shootouts ... Newman 14-9 Quinn 14-9 Myernick 10-8 Arena 12-10 Dir 15-13 Rongen 13-13 Fitzgerald 11-13 Bradley 5-6 Hankinson 6-10 Osiander 5-9 Zenga 4-8 Calloway 5-10 This looks an awful lot like random scatter.
(Dear god, so many new posters. It's a miracle.) Numerista, I did something similar in this thread, looking for between-season team correlations. I personally would expect a goalkeeper to have a lot more effect on shootout percentage than a coach. Nevertheless, I found the same thing, pretty much no discernable relationship.
I have no idea about whether or not home teams were more likely to win the shootout, and don't really want to spend the time it would take to find out. If someone wants to go through old match reports, here's the place to begin (or maybe Andy Mead or someone else knows a better way). I suppose my questions then are: 1. How much did they change how they played? and 2. How much do these changes in play affect the outcomes of games? Obviously these questions deserve their own thread, but I'd wager it's not as much as usually is believed.