Login

juke · 02-29-2020, 10:01 PM

2282 words, ready for grading

Hey guys, round 3 of my ongoing statistical look into the chances of wins for each team and match up in the playoffs. I’m going to give it almost no introduction, since I’ve done so twice already, so check out the following links if you haven’t read the logic behind the post, and where the numbers come from. Just like last week, I’m going to leave a google docs where you can actual see the probability of each scoring outcomes.

Round 1 write-up: https://simulationhockey.com/showthread.php?tid=103671
Round 2 write-up: https://simulationhockey.com/showthread.php?tid=103713

As always, I have to stress that these aren’t predictions: they are literally just percent chances that each team will win any given game or series, based on their team scoring relative to the other team. They are not my choices for who will win, nor as they carefully analyzed direct match ups between the two teams based on their team composition. In fact, during the last round of the playoffs, the team that the stats reported as favorites lost 3 out of the 4 series. In part because most of the match up predictions were just a few percentage points off from each other at the start of the series, and because it’s just a percent probability, not an actual predicted stat. So don’t take these stats too seriously, they’re just for fun. Honestly, at this point, if you took the opposite of what these stats said it might have more predictive power.

The only thing that’s changed in the code since the last round’s analysis is that I switched up how much I weighted current team scoring stats. Last round, the predictions were based on team scoring throughout the year, but they were very heavily weighted on stats from more recent weeks. The logic behind that was because certain teams would improve faster than others each week, depending on how many active players/rookies they had. Since last week the predictions didn’t do so great, I decided to weigh the more recent weeks less, and really let the season long stats shine. I say this knowing that now more recent stats will be more important again, just to spite me. But it doesn’t matter, we’re on to what the numbers say about each match up.

As always, I'll be updating each team's percent chance of winning, and how well they're matching up against each throughout the series after every game, so keep following along if you're finding the numbers interesting.

Scoring probability graphs: https://drive.google.com/file/d/1Xsr2ult...sp=sharing

St. Louis Scarecrows vs. Detroit Falcons Falcons

Pre-series stat breakdown:

[Image: matchup1_pre_stats.png]

Game by game breakdown:

[Image: matchup1_breakdown.png]

The pre-series stats, which are based on playoff stats for both teams, describe a showdown in which Detroit is favored in almost every single category. But I think that’s honestly skewed in the fact that St. Louis had a tough 6 game series against a high seeded Colorado team, where-as Detroit swept the Berserks in 4, so their stats may be inflated compared to St. Louis. But the fact of the matter is that Detroit leads the head to head in almost every stat other than penalty killing and hits. Most notably, St. Louis had a pretty rough powerplay in round 2, while Detroit’s powerplay went nuclear on Newfoundland. Speaking from experience, St. Louis will want to stay out of the box against Detroit, and they will want to capitalize on all of their special teams opportunities themselves. Games in Detroit have a 56.1% chance of a Falcons win, with an expected score of 3.79 to 3.38 to Detroit. However, the Scarecrows once again have a big swing in home ice advantage, giving them a 58.2% chance of winning in St. Louis. The expected outcome of those games being 4.19 to 3.21 Scarecrows. This gives the Falcons the overall chance of 55% of winning the series. These numbers are almost identical to last round for the Scarecrows, where they found themselves down early 0-2. But the Scarecrows had the advantage at home then too, and started their comeback run once they went home, and won all 3 games in St. Louis. So while Detroit is statistically the favorite, don’t be surprised if St. Louis repeats what they did last round here.

After game 1: Detroit kept their momentum from last round, taking home game one and still have yet to lose in the playoffs. However, as much as a game 1 loss stings, I think St. Louis will look at the stats of the game and take them with confidence that they can compete in this series. The Scarecrows had better numbers almost across the board compared to the Falcons: fewer PIM, more shot blocks, better special teams, better corsi. The Falcons managed to find the back of the net just one more time than the Scarecrows off of more elite shooting and better goaltending, but their powerplay finally came back down to earth and they now remember what it's like to go a normal 1 for 5 on it. The Falcons increased their chances of a series win to 63.8%, but the numbers indicate that St. Louis played more like the team that won 4 straight games in the previous round, rather than the team that lost the first 2. If they can hand the Falcons their first loss of the playoffs and split the series 1-1 in Detroit, they'll have a huge momentum swing headed to St. Louis, where they're favored.

After game 2: The Scarecrows owe Doyle for this game, stopping the disproportionate amount of shots faced to even the series up at 1 before heading back home. After going something crazy like 45% on the powerplay against us last round, the Falcons have gone cold on the an advantage for the second straight game, and they've now experienced something completely new for them in the playoffs: a loss. Once again, the Scarecrows seem like they're outplaying the Falcons in the stat by stat comparisons, and they now have a 54.7% chance of a series win. The number being higher than 50% is from the fact that they were more heavily favored at home, and now have home ice advantage in the series. Part of me thinks that it's just a random byproduct of Simon, but they did go 3-0 at home vs. 1-2 on the road last series, so they have a lot of confidence heading home.

After game 3: The back and forth action of this series continues, and it seems like each team is taking turns having a hot goalies carry them to victory. This time it was Jobin who saved an incredible 33/35, as the Falcons walk away with game 3 at road, the first loss St. Louis has had there in the playoffs. A huge factor in this game was the special teams: the Falcons made the most of their opportunities and went 2/3, while the Scarecrows fell back to their last round struggles and went 1/6. The series stats seem like two pretty evenly teams matched up, and I think whichever goalie can get a more sustained hot streak may win this series. As in stands after 3 games, Detroit has a 67.4% chance of the series win.

After game 4/5: I didn't have time to update after the game 4s, but with the Scarecrows' win, it temporarily moved them to a 47.8% chance of winning the series, and the game 5 win skyrocketed them to a 76.6% chance of closing out the series in one of the final 2 games. Looking at game five's stat sheet, it looked like the trend of better goalie continues to be true in this contest: Doyle was fantastic, stopping 31 of the 34 shots faced, while Jobin let in two more goals on the same number of attempts. Remarkably, the save percentages for each goalie have been dead even in the series long match up, but it seems that instead of close goalie battles in each game, the equality is due to each goalie taking turns are being unstoppable vs. swiss cheese. The special team's for St. Louis continues to be pretty bad for the second round in a row, but I think they'll take the worse special teams as a small inconvenience to have better team corsi, shooting percentage, shots blocked, hits, and of course the 3-2 series lead. They go home where they've been fantastic this offseason, to try and close this series out in game 6, just like they did last round.

After game 6: Haven't really been able to look into the final graphs of the series, but another great home win by St. Louis to move them on to the championship. Congrats to them, theyir underdog run faces possibly the hardest opponent yet in the red-hot Outlaws for the cup.

Outlaws

Anaheim Outlaws vs. Carolina Kraken Kraken

Pre-series stat breakdown:

[Image: matchup2_pre_stats.png]

Game by game breakdown:

[Image: matchup2_series_tracker.png]

I am so excited for this series. It’s remarkable how close the numbers predict this to be. And how good both of these teams were throughout the year. The pre-game match up stats say that both teams have equal advantages throughput a variety of categories, but the main stat that jumps out to me is that Carolina lets in a significantly more goals than Anaheim. However, the Kraken seem to have more lethal shooters, and are shooting at a much higher percentage than the Outlaws, so those stats themselves could even each other out in this particular match up. Running the predicted game outcomes was an even more incredible surprise highlighting how close this series could be. Astonishingly, both of the away teams have a 49.8% chance of winning on the road. The season scoring stats of both of these teams predict that basically any given game is a coin toss regardless of the location of the game. Speaking of coin tosses, the game outcomes are so close that neither team has an advantage in the series, and the percentages round up to 50% for each team. I hope that this series lives up to the insanely close predictions, and it should be a suspenseful affair. To be honest, I’m glad that the predictions favor neither team, because it’s impossible for it to be wrong, and this serious features two amazing teams that I wouldn’t want to have to pick one over the other anyways.

After game 1: For the second straight round, the higher seeded Kraken lose game 1 at home. This game was incredibly similar to the other game this round, in that while Anaheim came home with the win, the Kraken seemed to outplay them at every turn but just happened to get slightly worse goaltending and lost by 1. The Kraken had better team corsi, special teams, hits, and shots blocked, but at the end of the day all that matters is the 1 more goal that the Outlaws got. The Kraken got 2.5 times the amount of penalty minutes as the Outlaws, and while Anaheim couldn't capitalize, the Kraken will certainly want to tone that number down going forward in such a close series. The game 1 win moved Anaheim to a 65.6% chance of winning the series, however with any given game being a 50/50 chance for a win for either team, the Kraken can easily come right back to even those odds back to 50% with a win tonight.

After game 2: After their series win last round, Carolina is looking to replicate everything that led to their success, which includes losing the first 2 games at home. Game 3 was somewhat of a blow out - Anaheim threw way more shots on net than the Kraken,and ended up winning by 3, a score differential that was only about 7% likely to happen. Carolina certainly aren't panicking yet, as this is the exact same situation they found themselves last round. In fact, they have better odds of coming back from 0-2 than they did at this time last round. As it stands, the Outlaws have an 81.2% chance of closing out the series, which they'll try to do by continuing to get the edge in corsi, special teams, and stay even in goaltending.

After game 3: Anaheim continues to fill the Kraken's net, going back to back games with 5 goals, and now have a 3-0 series lead. Both teams played a pretty even game, but the big difference was not special teams, but penalties themselves. For whatever reason, Simon thought the Outlaws played a very disciplined game, as they weren't shorthanded a single time the entire game. Compare that to the Kraken, who Simon cursed with 7 penalties which the Outlaws scored on 2 of, which ended up being the final difference in the game score. Carolina has a huge hole to climb out of in the rest of the series, with only a 6.24% of completing the comeback and winning in 7. But hey, it's Simon, anything is possible.

After game 4: Well Anaheim just ran away with this series. I would like the thank the Kraken for making sure that us Berserkers aren't lonely in the dust club, as well as giving me less graphs to have to update on a week where I'm busier with work. Nothing about the match up stats really indicated that Anaheim dominated this series. Carolina's ultimate downfall seemed to be Simon giving them way more penalties than the Outlaws, and the Outlaw's far superior shooting percentage than the Kraken. As it stands now, Anaheim has to move into the discussion as the favorite next round, as they took care of business against a great Carolina team with ease, while the remaining two teams are duking it out and possibly going to 7.

Nhamlet · 02-29-2020, 10:15 PM

Will the birds be scared away? Find out on next episode of "Simon Says"

*thiefofcheese* · 02-29-2020, 10:30 PM

02-29-2020, 10:15 PMNhamlet Wrote: Will the birds be scared away? Find out on next episode of "Simon Says"

Simon is a dick

***Chevy*** · 02-29-2020, 10:33 PM

02-29-2020, 10:30 PMthiefofcheese Wrote:
02-29-2020, 10:15 PMNhamlet Wrote: Will the birds be scared away? Find out on next episode of "Simon Says"

Simon is a dick

Yes. Yes he is

juke · 03-01-2020, 09:11 PM

Sunday games updated

BDonini · 03-01-2020, 10:23 PM

juke · 03-03-2020, 10:17 AM

updated

***Chevy*** · 03-03-2020, 11:23 AM

Quote:Carolina's ultimate downfall seemed to be Simon giving them way more penalties than the Outlaws

Granted, Simon can be a fickle sim at times but there was a back and forth on just this idea yesterday in the index thread.

Anaheim has an average check rating of 43. Carolina has an average of 47, with more players at 50+ than Anaheim has. More checking can lead to more penalties.

Not an excuse, because even I agree that Simon puts out a ridiculous amount of penalties sometimes (Anaheim has been hit by it as well).

Just observing a data point.

LordBirdman · 03-03-2020, 11:27 AM

Juuuuuuuuuuuuuuuuuukkkkkkaaaaaa

juke · 03-03-2020, 11:44 AM

@Geekusoid Yeah that's a good point, a better way to phrase it shouldn't have been blaming Simon as much as just blaming penalties in general. In fact, I quickly looked at every team's average checking vs. PIM for the regular season, and small differences in average checking do make a pretty big difference in team PIM. On average, every 1 point increase in team checking led to 12 extra penalty minutes on the year, and that number would've been way higher if not for Vancouver and Anaheim being outliers with high PIM for pretty low checking. Maybe the real story is that Anaheim finally regressed closer to the real relationship between PIM and checking this series, while Carolina stayed the same as the regular season.

[Image: pim_data.png?width=412&height=338]

@LordBirdman Hi Birdman

EricNCSU · 03-03-2020, 12:36 PM

03-03-2020, 11:44 AMSmalinowski7 Wrote: @Geekusoid Yeah that's a good point, a better way to phrase it shouldn't have been blaming Simon as much as just blaming penalties in general. In fact, I quickly looked at every team's average checking vs. PIM for the regular season, and small differences in average checking do make a pretty big difference in team PIM. On average, every 1 point increase in team checking led to 12 extra penalty minutes on the year, and that number would've been way higher if not for Vancouver and Anaheim being outliers with high PIM for pretty low checking. Maybe the real story is that Anaheim finally regressed closer to the real relationship between PIM and checking this series, while Carolina stayed the same as the regular season.

@LordBirdman Hi Birdman

Fucking LPLL skewing that graph for the whole damn team... LOL

Navigation

Extra Menu

About us