nice reply, thanks It's kind of silly that margin of victory is not used... all you do in an algorithm is ignore "run up goals." Plus, in college soccer I have never know a school that did not "run up the score" The Same in international play. who calls off the dogs? off top though, thanks again for the comprehensive breakdown.
Agreed, it's easy to plug goals-scored into any algorithm. Other sports rating algorithms, even some of the external algorithms adopted by NCAA in the old BCS average-of-six computer polls, did consider margin. Human voters are free to do so (but subjectively, by "feel", not by any code-like procedure), even in today's CFP committee. Obviously, FIFA majors use GD as the 1st group tiebreaker. It's just a policy decision by NCAA that RPI, as used by them for selection shows, shall not bake scoring into the computation. It does make results reporting a tad easier. IIRC, one of their stated rationales is that student-athletes shall have greater concerns than tracking goal-differential over a full season. They know it makes RPI a bit coarse, but they buy that trade-off (and maybe tweak by hand for selection snubs ). ~~~~ Over in wvb(/mvb?), the unofficial Pablo ranking is entirely an algorithm determined by points-scored, ignoring match-result (which is usually correlated.) It can produce really, really skewed results compared to ranking (inertia, name recognition bias), e.g. throughout preseason WSU was unranked and not receiving votes but #10 in Pablo (later confirmed to be largely accurate -- but we regress since), while conversely #10 FSU is consistently ~40-ish Pablo (and finally received a data point just this weekend). As a predictor of future wins, Pablo is big-data-designed to outperform any other use of the same data, i.e. its coefficients are tweaked so that its rankings maximize the win-prediction%, at about 70%. Schools are free to brag about all 4 of their different rankings (and we do ). That's a fine use of US-ian ingenuity, math leetness, and spare CPU cycles, and we're all ultimately on the same side. But for official selection decisions, NCAA will stick to RPI (+ committee oversight). Prepping for that is like diligently doing our taxes: convoluted and arcane, but them's the rules, and we all know it going into ... 2017
As a matter of interest, in terms of simple correlations of ratings with game results, rating systems that incorporate goal differential perform just about the same as the RPI.
agree, but in such tight neck to neck.... it can change things thanks to all for the insightful comments
I watched the second half of the Clemson-FSU game yesterday, thinking there would be some exciting soccer, and it was completely dull. FSU sat back, Clemson kicked it around and tried futilely to penetrate FSU's good defense, and there was no action. Nothing. Two good teams and a terrible second half. And the thing is, I've seen a number of ACC halves like that this year--polished soccer with little real action in terms of good chances--and I'm trying to figure out how very good teams can play such boring games. I think the first half of the last UVA - Duke game was the same way--tedious. I hesitated before making this point, as I have a ton of respect for the ACC, but I saw that European Soccer Fan made the same point about the Clemson-FSU game, and I thought: I am not alone! I think one reason might be that the teams spend a lot of time pinging the ball around in the back while the opponent, for stretches, is content to sit back. Eight passes in the back, four passes in midfield, then the team with the ball will try something, fail, defending team will possess for a bit, back to the other team, rinse and repeat. It is like two good boxers who spend most of a round moving around the ring, watching each other warily, feinting, a few punches that don't connect, maybe a little flurry before the ball rings. Maybe the teams all know and respect each other too much. The tournament should be more interesting as there will be teams with different styles.
It wouldn't be shocking for NC State to get an at large selection and Virginia Tech not get one. In fact, I believe that's the likely scenario. Miami, on the other hand, is likely to end up with an ARPI rank around #68. I believe it will not be anywhere close to getting an at large selection.
Can't watch the game here so can only see the stat sheet. 1 shot each. ouch. Looks like both subbing heavily. Is that to find better combinations in attack or are they both using this as a warmup to get a lot of players some burn before the NCAAs?
http://www.seminoles.com/ViewArticle.dbml?DB_OEM_ID=32900&ATCLID=211278607 FSU prevails in an evenly fought game. Very few dangerous chances. What seedings do the ACC teams get? I read earlier in the thread where they had FSU pegged as a 4 seed. Does that still stand? I can't imagine us being less than a 3 seed.
I thought UNC dominated the second half--but could get no quality scoring chances out of their possession. FSU is very good defensively. Both seem poised to do well in the tourney.
That is absurd, Miami's RPI is almost 30 spots lower than VaTech with Miami at 67 and VaTech and 39. NC State sitting at 45. Miami has no shot, VaTech and NC State have about an equal shot depending on if the committee goes based on RPI or better results. Both VaTech and NC State should get in, but both should also be sweating it out because it is no guarantee. While NC State only has 1 loss outside of the top 20, the win vs UNC is the only good thing on their resume. Va Tech also has just one loss outside the 20 as well, but only a tie against Notre Dame and Clemson to show for. To make it that much harder, they didn't play in ACC so can't have that as a deciding factor either. Will be interesting, and hard in my mind to leave one out if the other gets in.
NC State RPI is at 40, same as last week (no clue where you got 45... back in September was the last time they were that low?) VT at 38, up from last week, they had been behind NC State, as I recall and Miami is at 66. I can't see VT getting the nod over NC State as NC State made ACC the tourney and VT did not while playing a tougher ACC schedule than did VT.
Going by the ARPI that cpthomas puts out. He has VT 39, NC State 45, Miami 67 http://woso-stats.tk/college/2016/index_arpi
That's right -- except that Chris Henderson maintain's that website, not me. When you're looking at statistical rating systems for as many as 333 teams, which is what we have (because Francis Marion isn't playing DI soccer any more), the difference between #39 and #45 isn't very meaningful. The rating difference between Virginia Tech and NC State is .0042. With that rating difference, at a neutral site, the better rated team statistically will win 43.1% of games, tie 16.9%, and lose 39.9%. On the other hand, the difference between #39 and #67 is more meaningful. The rating difference between Virginia Tech and Miami FL is .0242. With that rating difference, at a neutral site, the better rated team will win 56.6%, tie 14.4%, and lose 28.9%.
http://www.theacc.com/news/acc-wome...selected-for-the-ncaa-championship-11-07-2016 Well the ACC got 7 teams in. 2 #2 seeds and 4 #3 seeds and NCSt was unseeded. SEC got 2 #1 seeds. It seems in the recent past the ACC has outplayed their seedings. It will be hard to do with 6 seeded teams, but we'll see. I may be being biased, but I have a hard time believing UF and USCe would have the same records in the ACC.
Does anyone know if the games this weekend (specifically ACC teams) will be televised on the ACC Network, or other station?
They are young but experienced having played 7 of the current coaches poll top 15 teams (this will be their 8th) with more success than one would think just glancing at a W-L line. Frankly, I think this game is a little unfair to both teams to open the tourney. Minnesota won a top league and NC State has been ranked as high as 17 in one algorithm earlier in the year. Both earned a less challenging game to open.
With ND losing first rd, if Duke wins today they host 2nd/3rd rd games, right??! Also, just got word Racioppi BROKE her leg (she's been out 9 games). The list her "day to day", but she's not coming back this year
That's right. As the remaining seed in that portion of the bracket, Duke will host if they win today. That is, assuming their bid to host included more than just the first round game -- which I mention because I know of at least one occasion in the past where a team entitled to host the Rounds 2-3 games did not bid for them, intentionally, due to insufficient hotel rooms to accommodate both a football game and four women's soccer teams and their fans. I doubt very much that type of problem would apply to Duke.
Well Duke won handily today. And yes, we always bid to host other rds, regardless if we are the higher seed or not. Plenty of hotels here. One right across the street from the field! Go Blue Devils
I'm not sure anyone realizes this....but NC State's Kia Rankin was at the Duke game today, sitting in the stands with her mom!! In case you're wondering, State is currently playing in Minnesota! Not sure what's going on there. Possibly transferring?