USPC Results Thoughts
So before the test I posed a few questions. Now that full results are out, I thought I'd revisit them.
1. Will the test regain some of the easier/standard "15 pointers" that have gone missing in recent years? ... [Will this] shrink "the gap" which has not been trending in the right direction.
Yes. This question arose from what seemed to me to be a drop-off of scores on the high end, particularly last year, when only one score (mine) was over 300 and where the 10th place US solver had a 170. I'd quickly evaluate this year's test as having lots of 15 pointers by comparison, and the 10th place spot was a tie at 241. So a good trend in scores but I'll want another year of this kind of balance to be sure.
2. What will the effect of over a year of monthly internet tests at LMI and other sites have on this competition? The "old guard" generally don't compete in these tests, except for me. Meanwhile, some of the up and comers are getting frequent practice at online competition.... [W]ill [there] be some more shifting at the top and perhaps a second shake-up in two years in the US team makeup.
This question had two subtexts behind it. The first was tied to a fear that the crowded puzzle calendar might diminish the beauty of this test which used to be one of the few internet tests out there. As I stated after the competition I thought the puzzle quality was very high, and the addition of Serkan Yurekli to the set of authors is one reason for this as well as the continued solid contributions of the many other regulars.
The second subtext had to do with my dark horse pick for potential 4th US team member. I'd told Nick and some others earlier in the year that based on his monthly performances at LMI (typically in the teens), as well as suddenly playing Nikoli.com and croco-puzzle every day, that willwc was putting in the time to improve enough to make the team, and I expected he'd be very close to if not get the fourth spot. I figured the experience of competing at a very good WSC last year may have motivated his training in the way my competing at a very good WPC six years ago as a B-teamer motivated me to climb from (then) 11th to 1st in a year by putting in hundreds of hours of extra solving and speed training.
I unfortunately didn't put the money down officially this year on Will, or get to choose him for my USPC fantasy team, but that's because no one was laying any action or wanting to draft this year if they didn't get the first pick. But I'm excited to see Will make the team. I didn't get a very good chance to meet him in Philly as I was running around helping the event succeed most of the time, but I look forward to getting to know him better soon.
At the same time I congratulate him, I must admit that losing Roger, with many years of experience and several top 3 finishes, will be a clear short-term loss to the team and the biggest unknown at the WPC may be how well we solve together as a group of four when we now have two people with only one event's worth of experience, and then me and onigame.
Commiserations also go out to thedan for getting the numerical qualifying spot of 4th (as in 2008) but being beaten out by the exempt onigame who at least this time was 5th and not much lower down. But for 1 more square to get into partial bonus, Dan would have been on the team. As every year players are curious as to why there are exempt team members, the rationale as I understand it is that there is some value in maintaining some team consistency and to not put everything on a single test when someone could have a bad day. If a solver is in the top N in the world (typically N=8-10), then they should be on the US team. As we both made the playoffs last year in Poland, we both were returning this year to the team. While my competitive drive makes being exempt meaningless in terms of the focus I put into the USPC, I'm not sure it has the same effect on Wei-Hwa. I had little pity for Wei-Hwa's post here after the test that he had not read the rules carefully on the Dynasty Sudoku and wasted 10 minutes without knowing empty squares couldn't touch, particularly when I'd posted a practice one here 12+ hours before the test with a solution he could have at least looked at. But I'm guessing his prep for the test was a 30 minute review of the rules and making sure his pencils and erasers were ready. This is all I have time for before an LMI test, and I've read the rules there wrong myself too so it happens. Hence the exemption rule regarding bad days.
3. Can I be clean without -5 penalties, particularly on the counting puzzle if there is one? Will the person I perceive as my closest competition shake off his own mistake-prone trends and move up to second or even first this year?
No longer relevant, no, and a definitive yes. I've congratulated Palmer already but question three was about him, and also sprang out of question two. There was no doubt Palmer was also putting in the time to improve from last year's strong performance. In the last few LMI tests it was undeniable that either (A) I was getting slower or (B) Palmer was getting much faster as our times were getting near identical on tests I would describe, in the absence of seeing other scores, as being very strong performances by myself. While MellowMelon would describe his performances as good up to a few months ago, he always found himself losing time on puzzles like sudoku and such, and also losing points for answer entry errors. I've been answering all his questions and giving tips on sudoku and Palmer has taken the time to get better. I'm not sure I need to be giving him tips anymore, but I really look forward to seeing what he can bring at the full WPC level. I never have transitioned well from being a 2.5 hour solver to a 2.5 day solver. Taking a long view of things, my typical WPC experience can be described as ~1st/2nd after day 1, 4th-10th after day 2, ready to retire from puzzling after day 3. I've come very close only once, in 2008 when I was seconds from winning the WPC and holding all the titles (US/World x Sudoku/Puzzle Champion) and without playoffs I would have won all 4 titles. Now I hold none. But the first member of the next generation of US solvers has emerged and shown amazing form. How much better can he get, and can he deliver a playoff-level performance in Eger? We'll see.
4. And most importantly, what secrets will the "Reconstructed Rules" hold, if we ever get to see them?
This was a bit tongue and cheek, as I figured it was 50% likely to just be a page Nick hadn't reviewed in his updates and not that real changes were happening. But this was the most unexpected improvement in the test. For years I've argued for removing the penalties on mistakes that are excessively punative. I have not argued for this for my own ranking, as except for the counting puzzle I have never been wrong on a USPC entry allowing for a rare typo which earned credit. But many other solvers have. Last year Palmer's final score of a 225 included at least 15 points of penalty as I recall. I'm not sure if Roger similarly had mistakes, but 15 points would have changed the order on the podium.
But the second surprise was not the -5 rule changing but that my "creative" optimizer-like scoring of the counting puzzle was also coming in. This so-called "Thomas Rule" as one puzzle blogger referred to it actually did the most to unbreak part of the USPC, specifically having a puzzle that many solvers will avoid due to the likelihood of a -5 for spending time on it. The success rate of this puzzle is sometimes in the teens, with only 30% trying it. Without any penalty for doing it, and with a potential small reward for getting close to the answer if you do put in the time as opposed to a -5, participation improved on it. 30% (of 50%) earned 15 points as they would have before, but another 30% (of 50%) earned 5 points where they would have gotten -5 before (or just not tried the puzzle).
Did the rule changes help me? No. If anything under the old rules I would have avoided -5 on the counting puzzle until I had only it left and needed to solve it. With the 3 minutes I spent on it to earn nothing, I would have finished the only other puzzle I didn't have done, the Word Connection, for a 374. The change wouldn't have affected the final rankings, but at least made me look closer to the champion.
So that's it for another year of the USPC. Many interesting stories amidst the disappointment of not winning one more title. But I've got areas to improve. Not in puzzle solving, but in test approach. What I notice on LMI a lot, and for the first time on a USPC during the last 60 minutes of this test when I had so many puzzles to go, is that the egotistical approach of "I'm going to finish all these puzzles anyway, so let's just do this one now" sometimes fails pretty spectacularly. I don't game-plan order very much. On LMI I almost always go front to back. On the USPC I do the first print-out pages and the STD and then solve back to front. Neither optimizes my score given how much I know about every puzzle type. I need to stop this. Last year's WPC day 2 failure was in part from being tired and error-prone and sick. But in part from really horrible time management on the long morning round where I did not do the puzzles I was best at because I figured I'd get around to them later. There is always room to improve more. I'm going to need to to get comfortable going for a sixth title in 2012.
1. Will the test regain some of the easier/standard "15 pointers" that have gone missing in recent years? ... [Will this] shrink "the gap" which has not been trending in the right direction.
Yes. This question arose from what seemed to me to be a drop-off of scores on the high end, particularly last year, when only one score (mine) was over 300 and where the 10th place US solver had a 170. I'd quickly evaluate this year's test as having lots of 15 pointers by comparison, and the 10th place spot was a tie at 241. So a good trend in scores but I'll want another year of this kind of balance to be sure.
2. What will the effect of over a year of monthly internet tests at LMI and other sites have on this competition? The "old guard" generally don't compete in these tests, except for me. Meanwhile, some of the up and comers are getting frequent practice at online competition.... [W]ill [there] be some more shifting at the top and perhaps a second shake-up in two years in the US team makeup.
This question had two subtexts behind it. The first was tied to a fear that the crowded puzzle calendar might diminish the beauty of this test which used to be one of the few internet tests out there. As I stated after the competition I thought the puzzle quality was very high, and the addition of Serkan Yurekli to the set of authors is one reason for this as well as the continued solid contributions of the many other regulars.
The second subtext had to do with my dark horse pick for potential 4th US team member. I'd told Nick and some others earlier in the year that based on his monthly performances at LMI (typically in the teens), as well as suddenly playing Nikoli.com and croco-puzzle every day, that willwc was putting in the time to improve enough to make the team, and I expected he'd be very close to if not get the fourth spot. I figured the experience of competing at a very good WSC last year may have motivated his training in the way my competing at a very good WPC six years ago as a B-teamer motivated me to climb from (then) 11th to 1st in a year by putting in hundreds of hours of extra solving and speed training.
I unfortunately didn't put the money down officially this year on Will, or get to choose him for my USPC fantasy team, but that's because no one was laying any action or wanting to draft this year if they didn't get the first pick. But I'm excited to see Will make the team. I didn't get a very good chance to meet him in Philly as I was running around helping the event succeed most of the time, but I look forward to getting to know him better soon.
At the same time I congratulate him, I must admit that losing Roger, with many years of experience and several top 3 finishes, will be a clear short-term loss to the team and the biggest unknown at the WPC may be how well we solve together as a group of four when we now have two people with only one event's worth of experience, and then me and onigame.
Commiserations also go out to thedan for getting the numerical qualifying spot of 4th (as in 2008) but being beaten out by the exempt onigame who at least this time was 5th and not much lower down. But for 1 more square to get into partial bonus, Dan would have been on the team. As every year players are curious as to why there are exempt team members, the rationale as I understand it is that there is some value in maintaining some team consistency and to not put everything on a single test when someone could have a bad day. If a solver is in the top N in the world (typically N=8-10), then they should be on the US team. As we both made the playoffs last year in Poland, we both were returning this year to the team. While my competitive drive makes being exempt meaningless in terms of the focus I put into the USPC, I'm not sure it has the same effect on Wei-Hwa. I had little pity for Wei-Hwa's post here after the test that he had not read the rules carefully on the Dynasty Sudoku and wasted 10 minutes without knowing empty squares couldn't touch, particularly when I'd posted a practice one here 12+ hours before the test with a solution he could have at least looked at. But I'm guessing his prep for the test was a 30 minute review of the rules and making sure his pencils and erasers were ready. This is all I have time for before an LMI test, and I've read the rules there wrong myself too so it happens. Hence the exemption rule regarding bad days.
3. Can I be clean without -5 penalties, particularly on the counting puzzle if there is one? Will the person I perceive as my closest competition shake off his own mistake-prone trends and move up to second or even first this year?
No longer relevant, no, and a definitive yes. I've congratulated Palmer already but question three was about him, and also sprang out of question two. There was no doubt Palmer was also putting in the time to improve from last year's strong performance. In the last few LMI tests it was undeniable that either (A) I was getting slower or (B) Palmer was getting much faster as our times were getting near identical on tests I would describe, in the absence of seeing other scores, as being very strong performances by myself. While MellowMelon would describe his performances as good up to a few months ago, he always found himself losing time on puzzles like sudoku and such, and also losing points for answer entry errors. I've been answering all his questions and giving tips on sudoku and Palmer has taken the time to get better. I'm not sure I need to be giving him tips anymore, but I really look forward to seeing what he can bring at the full WPC level. I never have transitioned well from being a 2.5 hour solver to a 2.5 day solver. Taking a long view of things, my typical WPC experience can be described as ~1st/2nd after day 1, 4th-10th after day 2, ready to retire from puzzling after day 3. I've come very close only once, in 2008 when I was seconds from winning the WPC and holding all the titles (US/World x Sudoku/Puzzle Champion) and without playoffs I would have won all 4 titles. Now I hold none. But the first member of the next generation of US solvers has emerged and shown amazing form. How much better can he get, and can he deliver a playoff-level performance in Eger? We'll see.
4. And most importantly, what secrets will the "Reconstructed Rules" hold, if we ever get to see them?
This was a bit tongue and cheek, as I figured it was 50% likely to just be a page Nick hadn't reviewed in his updates and not that real changes were happening. But this was the most unexpected improvement in the test. For years I've argued for removing the penalties on mistakes that are excessively punative. I have not argued for this for my own ranking, as except for the counting puzzle I have never been wrong on a USPC entry allowing for a rare typo which earned credit. But many other solvers have. Last year Palmer's final score of a 225 included at least 15 points of penalty as I recall. I'm not sure if Roger similarly had mistakes, but 15 points would have changed the order on the podium.
But the second surprise was not the -5 rule changing but that my "creative" optimizer-like scoring of the counting puzzle was also coming in. This so-called "Thomas Rule" as one puzzle blogger referred to it actually did the most to unbreak part of the USPC, specifically having a puzzle that many solvers will avoid due to the likelihood of a -5 for spending time on it. The success rate of this puzzle is sometimes in the teens, with only 30% trying it. Without any penalty for doing it, and with a potential small reward for getting close to the answer if you do put in the time as opposed to a -5, participation improved on it. 30% (of 50%) earned 15 points as they would have before, but another 30% (of 50%) earned 5 points where they would have gotten -5 before (or just not tried the puzzle).
Did the rule changes help me? No. If anything under the old rules I would have avoided -5 on the counting puzzle until I had only it left and needed to solve it. With the 3 minutes I spent on it to earn nothing, I would have finished the only other puzzle I didn't have done, the Word Connection, for a 374. The change wouldn't have affected the final rankings, but at least made me look closer to the champion.
So that's it for another year of the USPC. Many interesting stories amidst the disappointment of not winning one more title. But I've got areas to improve. Not in puzzle solving, but in test approach. What I notice on LMI a lot, and for the first time on a USPC during the last 60 minutes of this test when I had so many puzzles to go, is that the egotistical approach of "I'm going to finish all these puzzles anyway, so let's just do this one now" sometimes fails pretty spectacularly. I don't game-plan order very much. On LMI I almost always go front to back. On the USPC I do the first print-out pages and the STD and then solve back to front. Neither optimizes my score given how much I know about every puzzle type. I need to stop this. Last year's WPC day 2 failure was in part from being tired and error-prone and sick. But in part from really horrible time management on the long morning round where I did not do the puzzles I was best at because I figured I'd get around to them later. There is always room to improve more. I'm going to need to to get comfortable going for a sixth title in 2012.