Time for my 5-game updated projections and predictions...
With how little talent this team has and how bad they are, are numbers REALLY that important to show how bad we are? Are these numbers in anyway going to change how a team plays?
Me running the numbers isn't going to change anything, no. But this thread started out as seeing how bad the sabres were in terms of comparing stats to other historically bad teams. I just expanded on that to noticing/determining the relationship between opponent strength and results and using that to predict just how bad we're going to be the rest of the season. Everyone knows we're bad, the numbers aren't going to convince people the team is better or worse than they see with their own eyes, and thats not really the point. If people took one thing away from this, I would hope it would be a realization that the sabres are what they are, and any percieved hot streaks/tank off and cold streaks/tank on stretches of games have more to do with playing easier or tougher opponents respectively than any other explanation. But like I said before, I'm doing this for fun, for my own sake, because I'm one of those strange people that actually find crunching numbers to be an enjoyable thing to do, and its how I can personally give purpose to following the team throughout the rest of the season. If some other people appreciate the info I'm sharing - and it seems at least someone does - then all the more reason to continue to post what I'm observing.
Anyways, here's the basic metrics for how we're projected to finish at the end of the season based on where we are after 50 games played:
Wins: 23.0 (-0.7)
ROW: 13.1 (+0.3)
Points: 57.4 (+0.9)
GF: 159.1 (+13.3)
GA: 236.2 (+8.4)
Our
opponents in games 46-50 have a combined average point total of 55.2 points as of today. Over the course of the 5 games, our opponents gained enough points to bump them up to our second easiest 5 game stretch (vs the easiest 5 game stretch before the 5 games were played). Even adjusting for today's point total, my predictions for GF and GA were way off for this set of games.
We scored a total of 17 GF this set - the sabres highest scoring 5-game set in the season so far. However,
we also gave up 19 GA during this set, which is our second worse GA in a 5-game set this season - the worst being 20 GA during games 16-20, our hardest stretch of the season (opponents have 73.6 average points today).
In terms of differential, we finished the set a -2, which while less than I predicted 5 games ago, it is within the margin of error for the diff curve now after accounting for today's point totals. GF and GA for the set, however, may need to be considered outliers, or a step change in performance, depending on whether we return to the projections next set or continue to be off the projection.
Looking ahead, I'm lumping the next 7 games before the olympic break together, rather than having a split set of 5 games - 2 before and 3 after the olympics. Not exactly sure if a 7 game set will exactly fit in with the data so far using 5 game sets, but I'll adjust later on if I need to. In those
next 7 games, our opponents have an average point total of 61.4 points today - a pretty tough stretch compared to other sets of games. Using the relationship based on our play this season against similar strength opponents, we are
expected to score 10.7 GF during the next 7 games while
allowing 19.8 GA, for a predicted
diff of -9.1.
Basically, anybody worried about the tank thinking we've been playing better and gaining ground... well, forgetaboutit. The cupcake chunk of our season is over. We've got a tough schedule coming up not just the next 7 games - the 5 games after the olympics are even tougher (63.8 average pts today). The last 5 games was supposed to be our easiest set of the season to date, and in my evaluation, we underperformed based on how we've played against similar strength opponents. Sure, we scored a lot of goals, but we had no business giving up as many goals as we did. This was a set of games we were supposed to be able to have a + diff, and we finished -2. Last time we played a set with a similar average points, we came away with 7 points. Last 5 games, we only got 4 points. If it weren't for the 3 goal win vs. the jackets, even diff would have been well below the margin of error. But that is why we look at chunks of games and not just individual games. There is lots of game to game variation that gets averaged out over time. Anyways, this should be considered nothing but a complete success for the tank, but entirely underperforming even for the lowly sabres standards.
I've added a points prediction to my data sheet. Much like GF/GA/diff, there is a clear correlation between points earned and average opponent points. Its not as solid as the goals correlations though, so its a lot more likely to be wrong - the error band is about +/-1.5 to 2 points per 5 game stretch. That means if I make a prediction right now, the margin of error for the prediction would be +/-5 points in either direction of my predicted point total for the end of the season (using root sum of squares of the error) - in other words, really not much of a tighter prediction than just using the basic projection like at the beginning of this post. But I'm doing it for fun so whatever.
Next 7 games, we should only gain about 3.4 points based on our play vs. similar strength opponents. Which really means, we'll probably only gain 3 or 4 points before the olympic break. The
5 games after the olympic break, being as tough as they are, we'll be over-performing if we get 2 points in those 5 games. As of today, the sabres have just 35 points on the year. Looking at the rest of our schedule and the average points of our opponents today in 5 game chunks,
we are only predicted to gain about 23 points the rest of the season if we continue to play as we have so far this year. That puts us
~58 points +/-5 points at the end of the season. Not surprisingly, this correlates well with the simple extrapolation of our points today after 50 games performed at the beginning of this post. But there you go... the basic projection and the more advanced prediction both arrive at basically the same conclusion through different means, so I guess thats saying something :shrug:
---
Down below is the data chart that shows the information I've been sharing. The data shown are the GF/GA/diff 5-game set data points through 50 games (so 10 data points so far). The x-axis (horizontal) is the 5 opponents average points, and the y-axis (vertical) is the 82-game GF/GA/diff pace for each 5-game set. Additionally,
sets under Rolston are the open, outlined only data points, and all the
solid data points are games under Nolan. The second set of data points from the left (just above 55 points on the horizontal axis) are the last 5 games, games 46-50. So, you see what I mean when I say GF and GA really deviated from the curves there, though diff is basically still on the curve or within the margin of error anyways.
The main reason I wanted to show this chart is to show conclusively that there has been
no change at all in how the team is playing since the start of the season. That also goes to say that, there is
no evidence in the data that Nolan is getting more tangible results out of the team than Rolston did. In terms of goal differential, they both follow the exact same curve. To conclude there is a difference, there would be a noticable step change between Rolstons games and Nolans games. You see the open data points to the far right of the graph - those are Rolston sets against tough opponents that Nolan hasn't come close to yet.
The 2 Rolston sets in the middle of the chart have results that are perfectly aligned with the results nolan achieved vs. similar strength opponents. Nolan has 4 sets - 20 games - easier than Rolstons easiest set, and the metrics were improved with the easier strength of opponents.
So, the improvement in the stats can be 100% explained simply by realizing that Rolston got the toughest stretches of games on the schedule, and lately we've had some real easy stretches - we play better against easier opponents. What a huge surprise, I know, but I just feel the obvious needs to be pointed out in this case. People are worried about the tank, worried we're going to pass Edmonton and Calgary at this rate - This rate will not continue. Unless there is a step change in team performance, we're getting back to the tank starting today. Look at the data points just above 61 on the horizontal axis - thats where our next 7 games are, thats how poorly we should expect to play. And the 5 games after the olympics are currently at 64 on the horizontal axis. The only step change we're going to see is potentially a change for the worst after the trade deadline. 58 points to finish the season is already pretty bad though, and I would be surprised if even Edm can manage to out tank that. We're still as bad as we were when we started the season, and I believe the next 12 games will show that.
That is not to say that I don't think Nolan has done great things as far as getting players giving more effort and playing harder and all. Myers has been great under Nolan and so have some others. Ennis and Stafford are playing well again. Girgs is developing. Lots of good things. Just that none of those good things are effecting the bottom line - which is exactly what the tank crowd wants, ultimately. To develop and see promise while still losing for a good pick. Thats pretty much exactly what is happening this year. But this thread began to see just how bad the sabres were in terms of stats, and as far as stats go, we're still one of if not the worst team in the modern era.
Here is the data I'm refering to:
