Tracking the Blues’ Stanley Cup Quest—LOL

[carter333167]

Figures have been updated--12/2.


1. Games Remaining—57 (Home Games—27; Away Games—30)
2. Available Points Remaining--114
3. Estimated Points Needed to get a Wildcard Slot—95
4. Points Earned to Date—21
5. Points Percentage Needed over Remaining Games to get a Wildcard Slot—.649
6. Sample Record that would get us to target—35-18-4


Upcoming games
1. Oilers
2. Jets
3. Canucks
4. Panthers
5. Avalanche
6. Flames
7. Oilers

Click to read more...

 

[Ted Hoffman]

You beat me to something like this. Suggestions:

* Drop remaining available points; it's meaningless with this many games to go.
* Definitely show remaining home/road games. That's important, especially come early January.
* I would change the target to 95, because that's where Dallas and Colorado currently project. [And, it's where the line was last year.] Leave it there until we get pretty good evidence that line has moved, as in "not until at least the halfway mark and then just nudge it every handful of games as needed.
* I might also post an idea of what kind of record is needed to get to 95 points. Like, 95 pts = 44-31-7, at this point we need to go 37-23-4.
* If you really want to get fancy, make a guess on home/road record and what we need to get to there. For example, if we need to go say 23-15-3 at home this season then that's 21-16-4 on the road; respectively, the rest of the way we'd need to go 18-10-2 at home, 19-13-2 on the road. But again, admittedly that's getting cute.

 

[TruBlu]

Thanks for doing this Carter. This puts a visual format to what I was talking about in the fire Yeo thread. When we look at how many games we have to win vs. how many we can lose, it shows why many are skeptical that it can be done. We may not be specifically mathematically eliminated until early February, but this current play is going to bring the players to a spot where they will know by mid December whether there's a bona fide shot at actually making the playoffs. Whoever wins the presidents trophy won't have the win percentage that we would need to secure a spot at that point. Yeo has to go now or I can only assume we are tanking.

 

[carter333167]

You beat me to something like this. Suggestions:

* Drop remaining available points; it's meaningless with this many games to go.
* Definitely show remaining home/road games. That's important, especially come early January.
* I would change the target to 95, because that's where Dallas and Colorado currently project. [And, it's where the line was last year.] Leave it there until we get pretty good evidence that line has moved, as in "not until at least the halfway mark and then just nudge it every handful of games as needed.
* I might also post an idea of what kind of record is needed to get to 95 points. Like, 95 pts = 44-31-7, at this point we need to go 37-23-4.
* If you really want to get fancy, make a guess on home/road record and what we need to get to there. For example, if we need to go say 23-15-3 at home this season then that's 21-16-4 on the road; respectively, the rest of the way we'd need to go 18-10-2 at home, 19-13-2 on the road. But again, admittedly that's getting cute.

Thanks! A bit busy today but I will keep incorporating changes to make this more useful---probably will make changes later in the day.

 

[carter333167]

Thanks for doing this Carter. This puts a visual format to what I was talking about in the fire Yeo thread. When we look at how many games we have to win vs. how many we can lose, it shows why many are skeptical that it can be done. We may not be specifically mathematically eliminated until early February, but this current play is going to bring the players to a spot where they will know by mid December whether there's a bona fide shot at actually making the playoffs. Whoever wins the presidents trophy won't have the win percentage that we would need to secure a spot at that point. Yeo has to go now or I can only assume we are tanking.

You bet. Basically, we need to get on a solid run almost immediately.

 

[TruBlu]

As a point of reference, the team's that have won the president's trophy the last 5 years have finished with point percentages between .689 to .720. Projecting that we go forward at .7 here on out (not going to happen), we'd finish the season with 107 points. Now take into account the next four games. A 1-3 record (what I project), then we drop to 103 points, only 8 points over last year's cutoff. It really doesn't take long to be out of contention. That's assuming we start to win at the rate of a president's trophy winner. I want this team to make the playoffs, but i just don't see it happening without a major change. I personally think the hole has been dug too deep already.

 

[carter333167]

Basically, Yeo or a new coach needs to kick this team into gear ASAP.

If we go .500 in points percentage over the next 10 games (a stretch that includes four games against the Jets and the Preds and that would end on Dec. 9) , these will be the revised targets:

1. Games Remaining—54
2. Available Points Remaining—108
3. Target Points Needed to get a Wildcard Slot—95
4. Points Earned to Date—27
5. Points Percentage Needed over Remaining Games to get a Wildcard Slot—.630
6. Sample Record that would get us to target—32-18-4

 

[TheDizee]

pointless to talk about playoffs. team wont come close to sniffing them with yeo.

we should be doing the math needed to obtain the best chance at #1 pick until yeo is fired

 

[Ranksu]

St. Louis Blues Playoff Chances - Sports Club Stats

 

[converseman54]

I've been a fan of this site for a couple of seasons

2019 NHL Post Season Probabilities

 

[Ranksu]

I've been a fan of this site for a couple of seasons

2019 NHL Post Season Probabilities

1% to Blues win Stanley Cup.

 

[Stealth JD]

they’re so f***ed. A top-three pick next year isn’t unthinkable.

 

[Ranksu]

they’re so f***ed. A top-three pick next year isn’t unthinkable.

Kaapo Kakko at eliteprospects.com

 

[TruBlu]

I think DA is going to stick with Yeo, simply because we don't have a snowball's chance in hell at making the playoffs. At least season tickets will be cheap next year.

 

[Ted Hoffman]

St. Louis Blues Playoff Chances - Sports Club Stats

I remember why I hated projections off this site.

[Yes, there's a rant. I've spoilered it for those who really want to see it. I also might be a little peeved by other stuff at the moment, so apologies in advance.]

Spoiler

Each time the league owner sends in new scores it simulates the rest of the season by randomly picking scores for each remaining game.

Randomly picking scores is statistical garbage. You have each team's GF/G and GA/G for both the current season and prior season, it's not difficult to come up with an expected GF/G and GA/G and adjust that going forward based on simulated results. That is what you should be using to simulate scores, not numbers picked out of the air. If you want to get really fancy, you could derive this with each team's roster and the individual players' expected GF (again, using current and past seasons) to get a more precise calculation, but that's getting into statistical theory. You could also do things like simulate roster movements (recalls, games missed, etc.) and adjust for who's playing in goal (do we really think the Blues are more likely to win with Allen or Johnson) and adjust for hot/cold streaks, but that's getting way into the details.

But, if you're trying to argue you've got a great model that's super reliable for projecting the rest of the season, it seems kind of stupid to not include those kinds of things into your model.

The weighted method takes the opponents record and home field advantage into account when randomly picking scores, so the better team is more likely to win.

See the point above. It's not difficult to break this down into home GF/G and away GF/G, but for the players it starts to run into credibility issues. But really, you should be using that measure instead of just W/L records to determine chances to win.

The 50/50 method gives each opponent an equal chance of winning each game. Both methods let an appropriate percent of games end in a tie or go into overtime in leagues where that matters.

An "appropriate percent". Oh, OK. It's the equivalent of saying "I need X games to go into OT, so ... this batch will do" without regard for how likely it is for any given team to play OT games. This is actually one of my lesser objections, because somehow OT games have to get determined - but I'd like for simulated results to let them happen, not just have the OT Fairy descend and choose select games for it.

It repeats this random playing out of the season million of times

What absolute crap. At the end of the day, you're trying to figure out the percentage chance of some event happening and be certain to within some threshold. That has much more to do with how reliable the model underlying your work is and the assumptions contained within that model than "ooh, I ran this eleventy million times!" If you need that for this model, either you have volatility in your results [which should cause you to re-evaluate your model] or you're wasting time with unnecessary work. Even at this point in the season, results should start settling within several thousand simulations but they're still subject to considerable error because of how much of the season remains [and all the unknowns not being contemplated in the model]. Focus on that, not running your model zillions of times in some false quest for accuracy.

To help flush out each team’s highest and lowest possible seeds, I force them to win or lose all their remaining games for a small percentage of the simulation runs.

Yeah, sure, someone might be able to reel off 61 wins in a row. The statistical likelihood of that is beyond incredibly small. Plus, it looks contrived and artificial to say "in 4 million simulations it was more likely for [team] to accumulate 22-69 points than it was to finish with 70-75 points." Really? Just say "here's the maximum/minimum points achievable" and simulate results using the assumptions going in. If someone really happens to reel off the remaining 61 games, fantastic - but let it happen via the model, don't artificially cram it in. Quit being cute trying to look ultra-accurate.

Short version: I think it's inaccurate until significantly less of the season remains and the chances for error decrease accordingly. About the time it starts to get accurate, you can derive the same results with considerably less effort than that site puts in.

 

[TruBlu]

I remember why I hated projections off this site.

[Yes, there's a rant. I've spoilered it for those who really want to see it. I also might be a little peeved by other stuff at the moment, so apologies in advance.]

Spoiler

Randomly picking scores is statistical garbage. You have each team's GF/G and GA/G for both the current season and prior season, it's not difficult to come up with an expected GF/G and GA/G and adjust that going forward based on simulated results. That is what you should be using to simulate scores, not numbers picked out of the air. If you want to get really fancy, you could derive this with each team's roster and the individual players' expected GF (again, using current and past seasons) to get a more precise calculation, but that's getting into statistical theory. You could also do things like simulate roster movements (recalls, games missed, etc.) and adjust for who's playing in goal (do we really think the Blues are more likely to win with Allen or Johnson) and adjust for hot/cold streaks, but that's getting way into the details.

But, if you're trying to argue you've got a great model that's super reliable for projecting the rest of the season, it seems kind of stupid to not include those kinds of things into your model.


See the point above. It's not difficult to break this down into home GF/G and away GF/G, but for the players it starts to run into credibility issues. But really, you should be using that measure instead of just W/L records to determine chances to win.


An "appropriate percent". Oh, OK. It's the equivalent of saying "I need X games to go into OT, so ... this batch will do" without regard for how likely it is for any given team to play OT games. This is actually one of my lesser objections, because somehow OT games have to get determined - but I'd like for simulated results to let them happen, not just have the OT Fairy descend and choose select games for it.


What absolute crap. At the end of the day, you're trying to figure out the percentage chance of some event happening and be certain to within some threshold. That has much more to do with how reliable the model underlying your work is and the assumptions contained within that model than "ooh, I ran this eleventy million times!" If you need that for this model, either you have volatility in your results [which should cause you to re-evaluate your model] or you're wasting time with unnecessary work. Even at this point in the season, results should start settling within several thousand simulations but they're still subject to considerable error because of how much of the season remains [and all the unknowns not being contemplated in the model]. Focus on that, not running your model zillions of times in some false quest for accuracy.


Yeah, sure, someone might be able to reel off 61 wins in a row. The statistical likelihood of that is beyond incredibly small. Plus, it looks contrived and artificial to say "in 4 million simulations it was more likely for [team] to accumulate 22-69 points than it was to finish with 70-75 points." Really? Just say "here's the maximum/minimum points achievable" and simulate results using the assumptions going in. If someone really happens to reel off the remaining 61 games, fantastic - but let it happen via the model, don't artificially cram it in. Quit being cute trying to look ultra-accurate.

Short version: I think it's inaccurate until significantly less of the season remains and the chances for error decrease accordingly. About the time it starts to get accurate, you can derive the same results with considerably less effort than that site puts in.

I use that website, too. I also use it closer to the end of the season. I remember many of us looking at it last year with 10 games left and saying there was no way we were going to be put out.

 

[Ranksu]

@Irish Blues I use it at spring time what it takes to get playoffs (point wise and possible playoff opponent)

I post it here, 'cus it maybe fits under this thread.

 

[TruBlu]

@Irish Blues I use it at spring time what it takes to get playoffs (point wise and possible playoff opponent)

I post it here, 'cus it maybe fits under this thread.

The numbers are good, but it doesn't give a good explanation of what is going on "right now." We are about 3 weeks aways from being eliminated (not mathematically)
realistically.

 

[AjaxManifesto]

We need a change!

 

[Ted Hoffman]

The numbers are good, but it doesn't give a good explanation of what is going on "right now." We are about 3 weeks aways from being eliminated (not mathematically)
realistically.

I think about that, and then I think about having to eat crow when the '08-09 team went on a run for the ages after sitting dead in the water at the midseason point with Kariya out, McDonald out, Brewer getting shelved for the season, EJ gone since training camp, Chris Mason as our starting goalie, and us having to rely on Backes (3rd year in the NHL), Perron (2nd year), Berglund (rookie) and Oshie (rookie) while being led by Boyes and Tkachuk with a defense of Colaiacovo, Jackman, Woywitka, Polak, McKee, and Weaver. There was no freaking way that team should have come back from 11 points out and dead last in the West on January 13 to put together a 21-7-6 run that landed us in the 8-spot on March 29, much less finished as the 6-seed.

It's why I won't say "there's no way they can make it back" but I will say "the chances of it happening are getting smaller" and "it's getting less likely."

 

[TruBlu]

I think about that, and then I think about having to eat crow when the '08-09 team went on a run for the ages after sitting dead in the water at the midseason point with Kariya out, McDonald out, Brewer getting shelved for the season, EJ gone since training camp, Chris Mason as our starting goalie, and us having to rely on Backes (3rd year in the NHL), Perron (2nd year), Berglund (rookie) and Oshie (rookie) while being led by Boyes and Tkachuk with a defense of Colaiacovo, Jackman, Woywitka, Polak, McKee, and Weaver. There was no freaking way that team should have come back from 11 points out and dead last in the West on January 13 to put together a 21-7-6 run that landed us in the 8-spot on March 29, much less finished as the 6-seed.

It's why I won't say "there's no way they can make it back" but I will say "the chances of it happening are getting smaller" and "it's getting less likely."

There's always a way, and I love the nostalgia, but we have a different situation today. The current coach doesn't fit the bill.

 

[HighNote]

December is one of the easiest months I've seen. 8 of the 13 games are against teams in the bottom half of the standings, and the other 5 aren't that special besides Winnipeg and Colorado. The Colorado game could be big (where have I said this before).

 

[carter333167]

December is one of the easiest months I've seen. 8 of the 13 games are against teams in the bottom half of the standings, and the other 5 aren't that special besides Winnipeg and Colorado. The Colorado game could be big (where have I said this before).

If the team still believes in itself and everyone is showing up every night to play for each other, we could go on a run. That's a big "if."

Stastny, in one of his last interviews as a Blue, said it best and most honestly: (paraphrasing) The league has so much parity now that, if even 2-3 guys take the night off, it is hard to win a hockey game.

 

[m1a2lt]

 

[carter333167]

Figures have been updated.