- My Forums
- Tiger Rant
- LSU Recruiting
- SEC Rant
- Saints Talk
- Pelicans Talk
- More Sports Board
- Coaching Changes
- Fantasy Sports
- Golf Board
- Soccer Board
- O-T Lounge
- Tech Board
- Home/Garden Board
- Outdoor Board
- Health/Fitness Board
- Movie/TV Board
- Book Board
- Music Board
- Political Talk
- Money Talk
- Fark Board
- Gaming Board
- Travel Board
- Food/Drink Board
- Ticket Exchange
- TD Help Board
Customize My Forums- View All Forums
- Show Left Links
- Topic Sort Options
- Trending Topics
- Recent Topics
- Active Topics
Started By
Message
re: ESPN's Football Power Index Predicts 9-3 Regular Season for LSU
Posted on 7/8/19 at 1:58 pm to Hold That Tiger 10
Posted on 7/8/19 at 1:58 pm to Hold That Tiger 10
quote:
How do you get a projection of 9-3 when LSU is favored in every game except one? Am I missing something?
Because none of those things are certain
1
Posted on 7/8/19 at 2:00 pm to toosleaux
quote:Oh?quote:You dumb
Salviati
Then show YOUR math.
Posted on 7/8/19 at 2:00 pm to Penrod
Yes, I know how that works. I’m not accounting percentages or anything over a total season, just a game by game standpoint. Ideally at 9.5 wins the projected total that would round up to 10 wins. So accounting a Alabama loss and throw in another game we shouldn’t lose puts us at a 10-2 record. I think it’s quite possible we go 11-1. One game at a time.
Posted on 7/8/19 at 2:01 pm to The Pirate King
quote:
Nah
Yep.
Otherwise Les Miles should not have been fired.
Posted on 7/8/19 at 2:03 pm to Diesel88
quote:
The FPI projection is 9.5 wins 2.8 losses
FPI projected us to play 12.3 games? I think you’re the one who botched the rounding.
Posted on 7/8/19 at 2:03 pm to Salviati
quote:
It's pretty simple. Add the percentages.
Just because LSU is favored to win all but one game does not mean that winning all but one is the statistically probable result.
For example, if LSU had an 70% chance to win 5 games and a 90% chance to win 5 games, it is statistically more likely to win 8 games than all 10 games even though it is favored in all 10 games.
.7 + .7 + .7 + .7 + .7 + .9 + .9 + .9 + .9 +.9 = 8
There's another thread dedicated to the lack of statistical knowledge of the posters on this board. They can't get it, but 9-3 is ESPN's FPI prediction.
Posted on 7/8/19 at 2:05 pm to TheTexasTiger7
quote:But it's not 9.5 wins. It's 9.357 wins.
Ideally at 9.5 wins the projected total that would round up to 10 wins.
quote:You're entitled to your prediction. I'm just posting what ESPN's FPI predicts.
So accounting a Alabama loss and throw in another game we shouldn’t lose puts us at a 10-2 record. I think it’s quite possible we go 11-1.
quote:Well, it's just a preseason prediction.
One game at a time.
Posted on 7/8/19 at 2:09 pm to TheTexasTiger7
It’s certainly possible we go 11-1, and might not even be that unlikely according to these projections. It just isn’t the record that the projections say is MOST likely.
And it’s (obviously) correct to say that FPI has us favored in 11 of 12 games. The only misread is to say that FPI predicts we go 11-1, which is what Salviati undoubtedly wanted to highlight with this thread. Those nonzero percentages of losses can’t just be discarded if we’re using the statistics properly.
And it’s (obviously) correct to say that FPI has us favored in 11 of 12 games. The only misread is to say that FPI predicts we go 11-1, which is what Salviati undoubtedly wanted to highlight with this thread. Those nonzero percentages of losses can’t just be discarded if we’re using the statistics properly.
Posted on 7/8/19 at 2:10 pm to Salviati
I’m sorry but 9-3 (unless there are major injuries) is completely unacceptable for this year. If this coaching staff can not at the very least achieve 10-2 with this team, then it will never be able to accomplish the goal of winning a national championship.
But let’s call these “percentages” what they really are. Made up bullshite that means nothing.
But let’s call these “percentages” what they really are. Made up bullshite that means nothing.
Posted on 7/8/19 at 2:12 pm to Salviati
29 downvotes. 0 upvotes.
Apparently, math is hard for some folks.
I didn't think it was possible to underestimate the ignorance of LSU fans.
Posted on 7/8/19 at 2:14 pm to J2thaROC
quote:
unless there are major injuries
Even then, it's next man up. That's why you recruit quality depth. Every team deals with injuries and the good ones are able to overcome them and win regardless. Heck, both teams in the 2018 national championship (Bama and UGA) dealt with significant injuries.
Posted on 7/8/19 at 2:15 pm to Salviati
If we accept the percentages of victory in ESPN's link then the odds of the following outcomes are:
0-12 0.0000000003%
1-11 0.0000002%
2-10 0.000021%
3-9 0.0008%
4-8 0.017%
5-7 0.2%
6-6 1.36%
7-5 5.9%
8-4 16.4%
9-3 28.5%
10-2 29.4%
11-1 15.5%
12-0 2.8%
You're welcome! As you can see the most likely outcome is 10-2, but barely ahead of 9-3.
0-12 0.0000000003%
1-11 0.0000002%
2-10 0.000021%
3-9 0.0008%
4-8 0.017%
5-7 0.2%
6-6 1.36%
7-5 5.9%
8-4 16.4%
9-3 28.5%
10-2 29.4%
11-1 15.5%
12-0 2.8%
You're welcome! As you can see the most likely outcome is 10-2, but barely ahead of 9-3.
This post was edited on 7/8/19 at 2:17 pm
Posted on 7/8/19 at 2:18 pm to Penrod
Nice tree results, but unnecessary to run a tree because, from a preseason viewpoint, the results of the games are more independent than dependent.
This post was edited on 7/8/19 at 2:22 pm
Posted on 7/8/19 at 2:22 pm to Salviati
FPI gave us a 99.9% chance of winning at A&M last year late in the 4th.
TIFWIW
TIFWIW
Posted on 7/8/19 at 2:26 pm to TheTexasTiger7
quote:
Yes, I know how that works. I’m not accounting percentages or anything over a total season, just a game by game standpoint.
Ah, so you are not doing it correctly, just doing it like a total imbecile instead?
But you definitely know how it works......
Posted on 7/8/19 at 2:30 pm to Salviati
quote:
Nice tree results, but unnecessary to run a tree because, from a preseason viewpoint, the results of the games are more independent than dependent.
I have no idea what you're talking about. However, it was necessary to do the calculations I did in order to know the odds of each of the season records given the odds of winning each game.
If you change the odds of winning a game my spreadsheet will give new odds on each season record.
If you want the spreadsheet post your email and I'll send it.
This post was edited on 7/8/19 at 2:31 pm
Posted on 7/8/19 at 2:30 pm to HotTakeHerald
I said I took it from a game by game standpoint. You’re an angry elf!
Posted on 7/8/19 at 2:31 pm to Salviati
quote:
Nice tree results, but unnecessary to run a tree because, from a preseason viewpoint, the results of the games are more independent than dependent.
It's not necessary to do anything. But the "tree" he posts shows what EV of wins cannot. If you are interested in LSU's chances of making playoffs or missing a bowl altogether, for example, EV won't give you your answer. You are much more interested in chance of having 0, 1, or more than 6 losses.
However, using today's ESPN predictions for the tree is problematic precisely because the results of the game become pretty dependent. Ex: We can say today that LSU is more likely to beat TAM on our last game of the year if we are undefeated going in than if we have 4 losses. Because it's more likely we have a better team.
Posted on 7/8/19 at 2:32 pm to Penrod
The lack of statistical understanding around here..
I actually started a thread on this exact subject (win probabilities vs. expected record) back in January and folks didn't seem to get it then either. People think the most probable outcome is to win all games when we are favored and win half the games when we aren't favored.
Just to clear this up for anyone who still doesn't get it - if your odds of winning each individual game are 75%, the most likely outcome is not that you will go undefeated.
To piggyback onto Penrod's very good analysis, here are the probabilities of various win totals if you accept the ESPN FPI odds:
Likelihood of 8+ wins: 93%
Likelihood of 9+ wins: 76%
Likelihood of 10+ wins: 47%
Likelihood of 11+ wins: 18%
Likelihood undefeated: 3%
I actually started a thread on this exact subject (win probabilities vs. expected record) back in January and folks didn't seem to get it then either. People think the most probable outcome is to win all games when we are favored and win half the games when we aren't favored.
Just to clear this up for anyone who still doesn't get it - if your odds of winning each individual game are 75%, the most likely outcome is not that you will go undefeated.
To piggyback onto Penrod's very good analysis, here are the probabilities of various win totals if you accept the ESPN FPI odds:
Likelihood of 8+ wins: 93%
Likelihood of 9+ wins: 76%
Likelihood of 10+ wins: 47%
Likelihood of 11+ wins: 18%
Likelihood undefeated: 3%
Posted on 7/8/19 at 2:34 pm to lostinbr
Page 2 of 5
Popular
Back to top
Follow TigerDroppings for LSU Football News