Script Tests

This page is mostly for the host but is here in case people are curious about it. Additionally, the script test information makes the News and Updates page look overwhelming, so they're on their own page.

There's countless script tests, but it's only the saved ones that appear below and will appear in order of occurrence:

Saved Test One

Dates: 24 - 30 Oct 2021

I play full games to test the script and to test my knowledge. I pose as two players to simulate multiple players and then flip coins to simulate game situations. I even roll dice to select categories and levels. The full parameters for these tests are too long to post here.
The top three categories of points earned are puzzles, questions, and jackpots, not necessarily in that order. Bonuses went under separate categories. My goal is that players earn most of their points through questions followed by puzzles. The purpose of the jackpot is to lower the disappointment of not knowing answers to questions, word scrambles, and whatnot. I have a feeling that the only way to move the jackpot out of the top three is to lower the amount of points that go there, but I don't want to decrease the number. I'm thinking about increasing it, but I digress.
In my head, I thought the points earned for streaks, both kinds, would be much lower than about 7,000 or five percent. I didn't realize that they add up. Of the two streaks, the overall streak was the higher number, which is what it should be.
I experimented with what would happen if I added a time bonus of one point a second remaining when asking for letters and the next category and level. It represented about 1.5, 2% of points earned. This isn't enough to really make a difference, but I hope this will motivate players to go fast.
I also experimented by changing the points going to the jackpot after a player solves a puzzle. Currently, that number is the difference between the full bonus and what the player earns, divided by four. I experimented with adding the full number and then changed it to divide by three. It moves winning the jackpot more towards the top earning category, so I'm not changing that bonus.
I also experimented by adding 100 points to the jackpot after asking each question. It does increase the percentage, but I don't think the jackpot will be the top earning category. At least, I hope not.

Saved Test Two

Dates: 6 - 21 Nov 2021

Actually, all games should be different, but I digress. Ideally, I want players to get most of their points from questions and puzzles. Unless I make some drastic changes to the jackpot, that is going to be in the top three. I would like for it to be number three, but it's possible to be number one and two. Anyway, here are some highlights of the second game's results:
The first game's victory margin, not including the bonus round, was about 32,000 points. When I added the bonus clock, the margin stayed the same, so it didn't have an effect on the game. However, when I added 100 points to the jackpot for each question asked, regardless of outcome, the margin decreased to about 29,500 points.
The second game's victory margin--once again, not including the bonus round--was about 51,000 points. This includes the bonus clock and the 100 points per question asked adjustments.
For one player, the top three scoring categories was puzzles, questions, and a jackpot. For the other player, the top three scoring categories was questions, a jackpot, and puzzles. The first player's top three weren't that far apart, percentage wise; however, the second player's top three had a big difference between the top two and puzzles. The first player solved most of the puzzles.
The scores for both players were neck and neck a few times during the game. The first player grabbed a big lead after solving the last puzzle and getting the final jackpot. In previous games, both players tend to score evenly, especially between questions and puzzles.
The second game had fewer streaks and a lot less points due to breaking the streak record. I wonder if it's because I missed more questions.
Both test games involved score corrections due to script errors. I hope I fixed the errors.
When I first script tested, I could solve puzzles fairly fast, probably because puzzles kept repeating. Now that I added more puzzles, they are repeating less, and I have trouble solving them. In fact, for some, I don't know the puzzle until the very last letter. This is probably a good thing. I'm saying this, because it's effecting the results of my analyses by giving results that will likely match a real game.
I'm not good with word scrambles. There are a lot of scrambles that I can't solve until the hint appears.

Saved Test Three

Dates: 28 Nov - 19 Dec 2021

The victory margin between the highest scorer and the lowest scorer was about 46,500. This number is after corrections and before the bonus round. (I test the bonus round, too, but I believe it works well, for the most part.) The difference between the highest scorer and the middle scorer was about 27,000, which is just above the halfway mark.
The top three methods of scoring for each player is as follows:

Highest: jackpot, questions, and puzzles at about 46.5%, 25.7%, and 12.2%.
Middle: questions, puzzles, and jackpot at about 46.1%, 16.2%, and 14.83%.
Lowest: questions, puzzles, and jackpot at about 66.9%, 23.6%, and 21.5%.

This game had some surprising results. The top three for each player doesn't match but is the top three I want. It looks like the highest player won thanks to the jackpot.
Winning streaks were virtually non-existent for this game. The main streak didn't go above three, and one player had an individual streak of seven. There was no bonus for breaking a streak, either.
That reminds me. I marked a question right that I got wrong. The question asked to put three events in order of occurrence. Well, I didn't realize at game time that I got the order wrong and, therefore, missed the question. Since this means I displayed the correct answer too soon, I had to toss the question. I subtracted the points given to the players and any points added to the jackpot. Additionally, I forgot that each player got one answer right. When I toss a question, players get points for all correct answers. I also had to adjust winning streaks, and it even required adjusting the winning streak bonus. There was one time in the game where a player got a main streak of three in a row and then another player got a streak breaker the next question. Unfortunately, tossing this question decreased that streak by one and negated those streaks.
Another mistake involved a puzzle. The script couldn't match a puzzle guess, because the apostrophe in the puzzle wasn't the standard, vertical apostrophe. The script doesn't remove the single quote characters that go left and right and even quotation marks that go left and right.
I messed up a score display for questions with more than one answer. The bonus points for getting all answers displayed properly, but it wasn't adding the correct number to the total points. I fixed that mistake at game time and made corrections. It took a bit.
One question was a level eight question, and I didn't realize it required three answers. (This is why I emphasize reading the question. Even I fail to read the question.) I only gave one answer and marked it right. Well, I noticed it after awarding points and a letter guess. I had to toss the question out and manually make adjustments, too many to list here. When typing this up, I forgot to award points for one correct answer, even though I tossed the question.
Players won the jackpot more times in this test, and the jackpot still had a high number of points after lowering the starting number to 1,000 per player.
For the last puzzle, the players revealed the entire puzzle, so the player that guessed the last letter got the jackpot. This makes it so there's no incentive for revealing all the letters.
When I script test, my objective is to go as fast as possible to get more questions, puzzles, and word scrambles. When the final puzzle got to the speed round, I changed my strategy, because I was tired and wanted to end the test. I delayed letter guesses as long as possible to run out the game clock. Not having that extra puzzle altered the results of this test.

Saved Test Four

Dates: 20 Feb - 6 Mar 2022

I made some adjustments to scores:

One player didn't get a bonus for time remaining on the bonus clock.
The script didn't calculate a winning streak bonus correctly. At game time, I thought it did it correctly.
I made a lot of reversals during the game, and I miscalculated them that gave a player an extra 50 points.
I gave partial credit to two players for their answers, but it somehow didn't add to their scores.

The script awarded the jackpot once. While there's no guarantee that the script awards the jackpot more than once, I want it to be more than once. I don't want a player to win the game by winning the jackpot. It may make the rest of the game seem redundant, and I can't fully imagine their frustration. For the latest script test, it was over 40,000 points, which made the player that got it the winner. I have to rethink the jackpot's algorithm.
The awarding of the jackpot once changed the top three of one player from questions (48.7%), letters (10.8%), and puzzles (9.1%) to jackpot (52.2%), questions (23.3%), and letters (5.2%).
The top three scoring for the other two players are as follows:

Middle Player: puzzles (40.1%), questions (30.3%), and letters (7%).
Lowest Player: questions (41.3%), puzzles (17.3%), and letters (15.6%).

I'm not good with scrambles. A player idled and lost control during the Puzzle Round test. I couldn't solve two scrambles. It was during Puzzle Round where a player had to have control before starting speed round on the next puzzle. That took almost three minutes off the clock. That was probably one less puzzle for that round.

Saved Test Five

Dates: 12 Mar - 3 Apr 2022

Once again, I had three players, and the jackpot bug affected the outcome. One player got a big lead.
This game had countless errors. I fixed most during the game, but I missed some:

I processed the pregame questions incorrectly, and one player didn't get their points.
I had troubles getting the script to add the correct number of points to the jackpot and displaying that. I caught most during the game, but I missed some and made adjustments after the game.
The biggest mistake this session is that I cut round three short. I did this test over multiple days, and I lost track of where I was during the game. I thought round three ended, but there was still 8:32 left on the puzzle clock. This is something I can't correct post game. I'm not going to toss the rest of the game to redo round three. During a live game, I hope to have the game at one session and not split over weeks like I do with my script tests. Fortunately, it didn't mess up Question Round and Puzzle Round in that order.
This game had a 69,000 swing between first place and last place with the middle player closer to the last place player. Here are the top three categories for each player:

First place: jackpot (46%), puzzle (21.5%), and questions (18.3%).
Second place: questions (37.9%), puzzles (25.7%), and letters (9.3%).
Third place: puzzles (44%), questions (34%), and letters (9.7%).
The problems with the jackpot skewed the top three. One player got a vast majority of those points.

The script awarded the jackpot five times. The first time was after the first question. I think the player got lucky. The next three were after the same question in round five. I don't know if that was an error in the script. The final one was after solving the last puzzle.

I goofed in my spreadsheet that shows the percentage of points players got in different categories so that requires adjusting previous script tests. I'll edit those directly instead of repeating it here.

Saved Test Six

Dates: 25 Apr - 30 Apr 2022

Once again, I had three players. Since this went over several days, but I got in a hurry to complete the test that I changed the parameters of the test, and I missed questions that I don't think I would normally. I hope to never have a live game when I'm tired. It may affect the outcome.
Out of curiosity, I counted the number of times I guessed vowels. It was a total of 39 with the highest scoring player requesting 14, the middle player requesting 13, and the lowest player requesting 12. This count only counts the number of times players lost points for vowels. There were a few times when players didn't lose points due to no consonants in the puzzle. Wow, that's a total of 9,750 points spent on vowels. It didn't affect the outcome of the game, but I was curious.
I fixed a lot of bugs, and I found more bugs. This test, however, had fewer problems where I had to stop the test to debug the script. I won't know for sure how long a game will actually take until having a live one. I estimate it at four hours, but with all the clock changes, it may be longer. I'll adjust it later if necessary.
For this test, the script and test parameters seemed to favor one player, who ended up winning the game. I don't know think it was due to script error. The final jackpot went to the player with the highest score. It looks like it chose randomly. No matter what I program in the script regarding randomness, there will be times when it seems it's favoring one player. There's nothing I can do about that as long as the random picker works correctly.
The jackpot's algorithm isn't quite right yet. The first jackpot got to about 50,000, and I would like it to award it at the 10,000-point mark. This number ignores the starting value. I will tweak with the algorithm again to hopefully award it sooner and more often.
First thing's first. It's time for post-game corrections:

For one question, it gave credit for one answer instead of two. That required adding 210 points, 70 of which was the all answer bonus.
It happened on two questions to the same player. The second time required adding 450 points, 150 of which was the all answer bonus.
There was a scramble as a question that got bonus points, and a player matched their guess. Unfortunately, it calculated the matching bonus before the other bonuses. That required a 15 point addition. I'll look into fixing it.
The jackpot didn't add the correct number of points for missed questions that required more than one correct answer. I manually adjusted the jackpot. I'm working on fixing the bug.
It looks like the addition of 100 points to the jackpot for each question regardless of outcome works. I'm not sure what to check in the code to verify that it works correctly.
I didn't program the addition of 1,500 points going to the jackpot for each round correctly. When the puzzle clock expires, 1,500 points goes to the jackpot (2,000 for Puzzle and Question rounds). It added the 1,500 points when the game clock expired, so I'll check that code.
The script awarded the jackpot twice. Both times required an adjustment: The first time deducted 200 points because of the 100 point error explained above. The second time deducted 1,700 points because of both bugs mentioned above this.
At first, the jackpot for incorrect questions wasn't adding the correct numbers. It was adding points for one answer instead of two or three, depending on the level. I fixed it later in the test, but in the process of fixing something else, the bug came back. I'm working on fixing it.

There was an approximate 118,000-point swing from the top scorer to the bottom scorer with the middle scorer about 14,000 above the bottom scorer.
The following is the top three scoring categories for each player, starting with the highest scoring player and ending with the lowest scoring player:

jackpots (52.6%), puzzles (21.9%), and questions (8.8%)
questions (34.4%), puzzles (33%), and letters (17.9%)
puzzles (35.5%), questions (32.2%), and letters (19%)

It looks like letters is fourth, and that's even after subtracting almost 10,000 total points amongst all players.

Script Test Seven

Dates: 7 May to 8 May 2022

First thing, I inadvertently skipped the choice for round two. Additionally, there was an error in the round numbers. A script error skipped round one. I thought I adjusted it back, but I changed the wrong variable.
Once again, I had three players with a slight change in one of the parameters. I tried to emulate players getting a question correct in part two by using dice rolls to determine if a player answers a question in part two. Since I'm playing these tests, it's technically cheating. I didn't like doing that. However, I slacked off on it towards the end of the test, but I think it wasn't due to guilt from cheating. I think I forgot about that change.
This test has a few corrections:

For reasons unknown, one player got double the points for partial credit. I'm not sure why that happened.
I don't know why, but a player got the wrong number of points for a question. The script didn't double the point value on a two answer question. I had to manually add points for one correct answer and the all answer bonus.
There was something wrong with allowing me to award points on a missed question for multi-answer questions. I had to manually award the points through score corrections.
On an educated guess question, I used the wrong command and awarded points for a correct answer instead of the reduced points. This also reduces that player's number of correct answers by one.

I made an error during the bonus round: On a right answer, I moved the level down instead of up. The player got the question right, so I made no corrections.
There was an approximate 58,000-point swing from the top scorer to the bottom scorer with the middle scorer about 3,000 above the bottom scorer.
I still had difficulty adjusting the script regarding selecting the next category and level. When the script is about to prompt a player to select a category and level, it erroneously thought there was already a selection made. I was removing the category and level variables at the wrong times. I finally got it to work correctly.
The following is the top three scoring categories for each player, starting with the highest scorer and ending with the lowest scorer:

jackpots (58.5%), questions (20%), and puzzles (6.7%)
questions (41%), letters (13.1%), and jackpot (20.1%)
questions (33%), puzzle (29.7%), and puzzle bonuses (9.6%)

Puzzle bonuses are points awarded for matching puzzle guesses and for time remaining.
The script awarded the jackpot six times. One player got it five of the six, which skewed the results.
The changes to the letter values may have changed the top three scoring categories, but it's too soon to tell.
The script tests emulate game situations through dice rolls and other parameters. The best way to tell for sure is to analyze a live game.
This test was the fastest I've done. The goal was to go as fast as possible to emulate how long a live game will take. I did stop to make some script adjustments at the beginning. That's how I will do live games. I'll make notes on adjustments and do them after the game unless the error is catastrophic enough to require fixing right away.

Script Test Eight

Date: 12 June 2022

I did an entire script test in one day. There were interruptions, and I took breaks, but it took one day.
There is one correction: I awarded points for an incorrect answer. Fortunately, it didn't affect the outcome of the game.
The objective of this test was to emulate its speed. I want to make sure it would take less than four hours. Unfortunately, I didn't time how long it took, so I don't fully know if it would take less than four hours.
The bonus round took about 12 minutes. That's too long, but I'm not sure on making adjustments, because I didn't think too long if I didn't know the answer to a question.
The clock stops a little too frequently. I'll have to think about how I can adjust it.
The script awarded the jackpot five times with one time being rather high. I believe it affected the outcome of the game to the point it probably made that player the winner.
There's a 31,500 point swing between first and second place. Here are the top three scoring categories for each player, starting with the highest scorer and ending with the lowest scorer:

jackpots (43.6%), questions (19.5%), and puzzles (11.8%)
jackpots (35.1%), questions (28.3%), and puzzles (15.1%)
questions (29.9%), puzzles (18.6%), and jackpots (18.4%)

I didn't count how many times players revealed the entire puzzle by guessing all its letters. It's difficult to tell if the jackpot affected the game. Each player got at least one jackpot.
I'm liking the newer system for vowels. The players guessed vowels 38 times for 19,000 points. I didn't track how many points they got back, and I didn't count the number of times the players guessed vowels with no consonants in the puzzle. (Remember that guessing vowels when there are no consonants left in the puzzle costs no points.) The goal is to discourage their use. Unfortunately, it's difficult to solve puzzles without them.
There are still some bugs in the script, and I'm not sure some features work the way I want them to, but the bugs weren't drastic to where I needed to stop the game to fix.