Bowhunting Contest Stats
02-21-2007, 07:12 AM
#21
Giant Nontypical
Join Date: Dec 2006
Location: Heaven is my home, temporarily residing in WNY :)
Posts: 6,679
RE: Bowhunting Contest Stats
Andy,
All I can say is WOW!!!! Awesome stuff! Tnhere is much to learn in your stats!!! This is magazine material Bro! You really need to look into that .... Congrats on a job ....more than well done!!!
All I can say is WOW!!!! Awesome stuff! Tnhere is much to learn in your stats!!! This is magazine material Bro! You really need to look into that .... Congrats on a job ....more than well done!!!
02-21-2007, 02:24 PM
#22
Typical Buck
Thread Starter
Join Date: Sep 2003
Location: Nanjemoy, MD
Posts: 998
RE: Bowhunting Contest Stats
And more than you really wanted to know... -fsh
The Consequences ofTeam Size
If you took all 410 participants of the 2006 contest,put the names in a hat, randomly drew out 41 teams, scored the teams, and repeated this cyle thousands of times,there would be an expected"frequency distribution".Every "draw" there would likely be a few teams with just one buck, and once in a while, a team would have 7 or 8 bucks. Once in a rare while, all teams would have only 2 or 3 bucks - the average being 2.6 bucks. This chart (histogram)compares the expected success of 10-person teams on average vs what actually occured:
Not a bad match, it looks reasonable. In fact, one way to run the contest is todraw names after the season is over(I want Lady Forge on my team!, lol). But this process would ignore the comraderie and encouragement a team fosters - maybe a few more folks got out there on a cold morning than they mighthave done otherwise.Team membership could have an influence on individual hunter success, but we'll have to conveniently ignore that issue and just treat hunter success as being independent of team membership or team size.
The size of the individual teams caninfluence the long term outcome of the contest as pointed out in my last thread. The "optimum" size of teams for the contest depends on who determines what constitutes "optimum".Statisticscan reveal some other issues related toteam size.
Suppose in 2006, HNI had used different sized teams? What are the consequences of team size? What if the team sizes were 1, 5, 10, 20, or 205? Sometimes it pays to look at extremes just to get a feel of what is happening.
Start with a team size of 1 (410 "teams" - the standardbig buck contest) - no comraderie there, just braggin' rights. Individual "team" scores willhavethe maxiumum difference from the mean score of all the bucks entered into the contest. (e.g. Lady Forge's whopper 190+buck was more than 90 inches greater than the ~100 inch contest mean). At the other extreme, a team size of205 (2 teams),would result in each team with ~ the same number of bucks, and~same averagescore; if you comparedthe two team's average buck scores, they would bevery close to the contest mean of all bucks entered.You might as well flip a cointhan have a two-team contest. So a team size bigger than 1 and less than 205 would be a good thing.
The size of the team (and given average hunter success of 0.266) will determine the frequency distribution of entries(as in the above chart for a team size of 10). It turnsoutonce a team getsto~10 entries and above, their average buck scorewill come close to the sample(contest)mean. In other words, if you construct a team size big enough (say 20 hunters/team)so thatit's likely some teams will have 10 or more entries, the average score of these teams'bucks will be close to the contest mean score -any "booners" will get lost in the noise by being averaged with their teammates'smaller bucks.In my earlier post withthe average score for teams posting at least 2, 3, or 4 bucks, note thatTeam 17's (8) entries averaged ~99, this is very close tothe averagescore of all bucks entered (~100). This is not a just a coincidence but is expected. (Even with a team size of 10 bowhunters, there was not a lot of difference in average scoresamong the top scoring teams).The larger the team size, the more bucks get posted, and the closer all teams' averageswill come to the contest mean. (Added in edit: one way to see this statistical trend is to compare the average scores of teams with at least 2, 3, and 4 entries - as the number of entries increase, note that the team averages come closer to the contest mean). So, given biggerteams, allshould have ~thesamescoring average, and harvesting the extra buck or two will makeall the difference between winning and losing (statistically speaking).
With smaller teams (say 6 or fewermembers) and accordingly fewer entries per team,the team averages will have a better statistical chance to deviate from the overall contest average. Now the performance of individuals of the team can carry the team, and the lucky teams with trophy bucks will stand out. Since trophy deer do not come along to us mere mortals but so often, one would expectthat the winning team of 6 hunters to have difficulty hanging on to their crown.As noted earlier, I expect Team 17 or their successors to hang in there for a while.
Just food for thought...
The Consequences ofTeam Size
If you took all 410 participants of the 2006 contest,put the names in a hat, randomly drew out 41 teams, scored the teams, and repeated this cyle thousands of times,there would be an expected"frequency distribution".Every "draw" there would likely be a few teams with just one buck, and once in a while, a team would have 7 or 8 bucks. Once in a rare while, all teams would have only 2 or 3 bucks - the average being 2.6 bucks. This chart (histogram)compares the expected success of 10-person teams on average vs what actually occured:
Not a bad match, it looks reasonable. In fact, one way to run the contest is todraw names after the season is over(I want Lady Forge on my team!, lol). But this process would ignore the comraderie and encouragement a team fosters - maybe a few more folks got out there on a cold morning than they mighthave done otherwise.Team membership could have an influence on individual hunter success, but we'll have to conveniently ignore that issue and just treat hunter success as being independent of team membership or team size.
The size of the individual teams caninfluence the long term outcome of the contest as pointed out in my last thread. The "optimum" size of teams for the contest depends on who determines what constitutes "optimum".Statisticscan reveal some other issues related toteam size.
Suppose in 2006, HNI had used different sized teams? What are the consequences of team size? What if the team sizes were 1, 5, 10, 20, or 205? Sometimes it pays to look at extremes just to get a feel of what is happening.
Start with a team size of 1 (410 "teams" - the standardbig buck contest) - no comraderie there, just braggin' rights. Individual "team" scores willhavethe maxiumum difference from the mean score of all the bucks entered into the contest. (e.g. Lady Forge's whopper 190+buck was more than 90 inches greater than the ~100 inch contest mean). At the other extreme, a team size of205 (2 teams),would result in each team with ~ the same number of bucks, and~same averagescore; if you comparedthe two team's average buck scores, they would bevery close to the contest mean of all bucks entered.You might as well flip a cointhan have a two-team contest. So a team size bigger than 1 and less than 205 would be a good thing.
The size of the team (and given average hunter success of 0.266) will determine the frequency distribution of entries(as in the above chart for a team size of 10). It turnsoutonce a team getsto~10 entries and above, their average buck scorewill come close to the sample(contest)mean. In other words, if you construct a team size big enough (say 20 hunters/team)so thatit's likely some teams will have 10 or more entries, the average score of these teams'bucks will be close to the contest mean score -any "booners" will get lost in the noise by being averaged with their teammates'smaller bucks.In my earlier post withthe average score for teams posting at least 2, 3, or 4 bucks, note thatTeam 17's (8) entries averaged ~99, this is very close tothe averagescore of all bucks entered (~100). This is not a just a coincidence but is expected. (Even with a team size of 10 bowhunters, there was not a lot of difference in average scoresamong the top scoring teams).The larger the team size, the more bucks get posted, and the closer all teams' averageswill come to the contest mean. (Added in edit: one way to see this statistical trend is to compare the average scores of teams with at least 2, 3, and 4 entries - as the number of entries increase, note that the team averages come closer to the contest mean). So, given biggerteams, allshould have ~thesamescoring average, and harvesting the extra buck or two will makeall the difference between winning and losing (statistically speaking).
With smaller teams (say 6 or fewermembers) and accordingly fewer entries per team,the team averages will have a better statistical chance to deviate from the overall contest average. Now the performance of individuals of the team can carry the team, and the lucky teams with trophy bucks will stand out. Since trophy deer do not come along to us mere mortals but so often, one would expectthat the winning team of 6 hunters to have difficulty hanging on to their crown.As noted earlier, I expect Team 17 or their successors to hang in there for a while.
Just food for thought...
02-21-2007, 07:21 PM
#24
Typical Buck
Thread Starter
Join Date: Sep 2003
Location: Nanjemoy, MD
Posts: 998
RE: Bowhunting Contest Stats
rybohunter
- Well, the obvious answer to your question is because Team XXfailed to repeat and Team 17 still hasa chance. Team XX was odds on to repeat (after all you guys did come in second out of 41 teams!), but whether the loss was due to skill ora statistical fluke, you were still beat out. The nature of the contest is such that both teams should take pride in their accomplishments.There is nothing in the "rules" that says you and your fellowTeam XX memberscan't informally keep track of your aggregate scoreas a "phantom" team relative toTeam 17 if that is your want -- even while you and yourTeammates are distributed amongst thenew2007 teams.It would be interesting if the top two "phantom team" scores (Team XX and Team 12) were kept as a point of discussion. But hey, this contest is for FUN and camraderie,in as much as mystat analysis is for fun too.Team 17 may not repeat, there is the chance that some new team can beat them, but the odds will below with justthe ~50 (??) teams competing next year.
-fsh
- Well, the obvious answer to your question is because Team XXfailed to repeat and Team 17 still hasa chance. Team XX was odds on to repeat (after all you guys did come in second out of 41 teams!), but whether the loss was due to skill ora statistical fluke, you were still beat out. The nature of the contest is such that both teams should take pride in their accomplishments.There is nothing in the "rules" that says you and your fellowTeam XX memberscan't informally keep track of your aggregate scoreas a "phantom" team relative toTeam 17 if that is your want -- even while you and yourTeammates are distributed amongst thenew2007 teams.It would be interesting if the top two "phantom team" scores (Team XX and Team 12) were kept as a point of discussion. But hey, this contest is for FUN and camraderie,in as much as mystat analysis is for fun too.Team 17 may not repeat, there is the chance that some new team can beat them, but the odds will below with justthe ~50 (??) teams competing next year.
-fsh
02-21-2007, 07:34 PM
#26
Giant Nontypical
Join Date: Aug 2003
Location: The OH-IO
Posts: 7,103
RE: Bowhunting Contest Stats
Team 17 may not repeat, there is the chance that some new team can beat them, but the odds will below with with justthe ~50 (??) teams competing next year.
BTW of the three years we have had the contest I have beenon teams that entered: in 20041 buck, 20052 bucks and 2006 1 buck.....
What are my odds of having a kick arse team next year?
02-21-2007, 09:21 PM
#27
Typical Buck
Thread Starter
Join Date: Sep 2003
Location: Nanjemoy, MD
Posts: 998
RE: Bowhunting Contest Stats
buckeyebuckhntr
Well, having one year's data does not negate the utility of statistical estimations, so you and I will have to disagree that my analysis has flaws on that basis.
One way to deal with limited data is to use "confidence intervals" about a statistical estimate, such as hunter success rate. (I hestitated to discuss confidence intervals in earlier posts, you have forced my hand). From the data we have, we can say that we are 80% confident that the average hunter success in the contest (mean = 0.266) lies between 0.23 and 0.30. In other words, there is only a 20% chance that the true average hunter success lies outside this interval. We can also say with the same confidence thatTeam17's true average hunter success(mean = 0.8) lies between 0.49 and 0.96. Team 17 true success rate is statistically likely well above the contest average. So it is improper to assign "26%" to them, and they are odds on to repeat their success or at least be in the running in as much as Team XX was this year. Anyway, time will tell.
Well, if you happened to be joined up withAntler Eater, Lady Forge, kshunter, and a few other skilledhunters onthe same team next year,your teamcould kick some arse. But in the meantime, just enjoy the contest and needle your teammates to spend more time in the stand, lol.
-fsh
Well, having one year's data does not negate the utility of statistical estimations, so you and I will have to disagree that my analysis has flaws on that basis.
One way to deal with limited data is to use "confidence intervals" about a statistical estimate, such as hunter success rate. (I hestitated to discuss confidence intervals in earlier posts, you have forced my hand). From the data we have, we can say that we are 80% confident that the average hunter success in the contest (mean = 0.266) lies between 0.23 and 0.30. In other words, there is only a 20% chance that the true average hunter success lies outside this interval. We can also say with the same confidence thatTeam17's true average hunter success(mean = 0.8) lies between 0.49 and 0.96. Team 17 true success rate is statistically likely well above the contest average. So it is improper to assign "26%" to them, and they are odds on to repeat their success or at least be in the running in as much as Team XX was this year. Anyway, time will tell.
Well, if you happened to be joined up withAntler Eater, Lady Forge, kshunter, and a few other skilledhunters onthe same team next year,your teamcould kick some arse. But in the meantime, just enjoy the contest and needle your teammates to spend more time in the stand, lol.
-fsh
02-21-2007, 09:31 PM
#28
Giant Nontypical
Join Date: Aug 2003
Location: The OH-IO
Posts: 7,103
RE: Bowhunting Contest Stats
FSH
I am by no means a math genius and I will have to take your word on these statistical estimations [8D]
My main point for asking you (about team 17's chances) is Iwas onwhat was originally preceived to be a "stacked" team in 05 but only 2 bucks were put down by the team.Most of the members of that team have fared rather well in the other years. Just wasn't our year so to speak.
[/align]
02-21-2007, 09:36 PM
#29
Typical Buck
Thread Starter
Join Date: Sep 2003
Location: Nanjemoy, MD
Posts: 998
RE: Bowhunting Contest Stats
buckeyebuckhntr - understood. & bad years happen to all of us - tougher when it happens to all your team mates in the same year too. So if you took all your former team mates' scores and constituted a "phantom" team for the 2006 contest, how would the team have done? -fsh
02-21-2007, 10:06 PM
#30
Giant Nontypical
Join Date: Aug 2003
Location: The OH-IO
Posts: 7,103
RE: Bowhunting Contest Stats
Of my 10 team mates of 05 only 4 of them were in this years contest including myself. Plus one other who didn't join this years contest but did take a buck.
So, of the ones I could locate for the 2006 contest 3out of the 5 scored bucks for 411 1/4inches. (Double Creek 136 7/8's, Ridgetop 136 5/8'sand Myself 137 6/8's).
Can't wait to see who I get teamed up with this year!
So, of the ones I could locate for the 2006 contest 3out of the 5 scored bucks for 411 1/4inches. (Double Creek 136 7/8's, Ridgetop 136 5/8'sand Myself 137 6/8's).
Can't wait to see who I get teamed up with this year!