Solo Shmup League: Rules, Guidelines & Game Choice Thread

[heisenbergman]

Looks solid. On overall scoring though, I think simpler might be better. Why not just use a running tab of every weekly score you've ever earned. It's not as important as the weekly score anyhow, it's just there to see how many points you have collected over the life of the league relative to other players.

I was hoping to have some sort of gauge of "season/career performance" if this is able to go on long enough --- and to be able to rank players according to that. Like in most pro sports for example, players have scores for the game, averages for the season and for their entire career.

Also what are the limitations for games we can nominate? I feel like we ought to at least be able to nominate console games seeing as we are voting. If no one wants to play them or they don't have access to them then that will be reflected in the votes I think. We could always just take a vote to see how many people want console games to be nominate-able versus how many don't? Maybe make a limitation of only 1 console nomination allowed (of your 3 nominations) in order to keep Mame options open. Just throwing out ideas, I hate to see this limited to only Mame.

Considering multiple people will be able to nominate games and everyone else will be able to vote on them, I was thinking maybe it's okay not to have limitations on nominations since, like you mentioned, interest in the game will be reflected in the voting anyway.

[trap15]

One concern I have with this is that posting a score could decrease one's average score and overall ranking. Scoring system should encourage participation and posting scores in subsequent weeks, not discourage it.

A possibility is to "deprecate" scores over time, so only the X most recent weeks apply.

[heisenbergman]

Overall ranking will be computed using the following formula:

• • 

• • W = player's overall score for ranking
  • R = player's average score across all weeks/games they've participated in
  • v = number of weeks/games the player has participated in
  • m = minimum weeks/games required to be ranked (we can set this to 1 for the meantime)
  • C = the mean average score across all league participants

One concern I have with this is that posting a score could decrease one's average score and overall ranking. Scoring system should encourage participation and posting scores in subsequent weeks, not discourage it.

Actually, I think not posting scores/participating will definitely decrease one's weighted average since the number of games participated in factors into the formula, and the smaller it is relative to other players, the lower one's score becomes.

Take for example a PlayerX with a score of 90, but has participated in only 3 weeks versus a PlayerY who has an average of 83, but has participated in 6 weeks... if we use C=60 and m=3, we have:

PlayerX = (90(3) + 60(3))/(3 + 3) = 75.00
PlayerY = (83(6) + 60(3))/(6 + 3) = 75.33

So PlayerY still has the higher weighted rating even with the lower average since his average came from more games compared to PlayerX.

I think this way, players can still afford to miss a couple of weeks without having a devastating impact on their weighted average; yet at the same time, it still encourages participation to ensure that you won't significantly lag behind.

[trap15]

The thing is that it means it's not worthwhile to post a score that would bring down your score average.

[heisenbergman]

The thing is that it means it's not worthwhile to post a score that would bring down your score average.

I think it would be a bit tricky to tell though how much of an impact posting one's score would have to one's average versus not posting the score.

One would have to:
(1) hold off from posting a score until the last moment so they could compute how much points they would get
(2) compute their points as a percentage of the highest scorer
(3) determine their ranking among all participants to be able to get their final points for the week [which would mean going through the whole thread to collate the scores of the other players]
(4) compute how much of an effect those points would have on their weighted average [which would mean they have to keep track of the median score across all players to use for the formula]
(5) and then do the same for all other players to see how much your rank will change as an effect of either posting or not posting your score

I don't know how competitive some of the folks here are to be willing to go through that just to get an upper hand, but I just think that that's such of an effort to go through just to ensure that you're not bringing down your weighted average for the week, and quite against the spirit of friendly competition as well. You'd have to practically duplicate what I'll be doing in gathering and crunching all the numbers just to see how much of an impact either posting or not posting the score will have to your overall ranking.

And even if one's score is so dismal that posting it is out of the question, the mere fact of it not being posted for that week will lower one's weighted average to some degree. And the more scores one doesn't post, the more the weighted average will decrease.

Ultimately... I think there is no perfect scoring system. Each will have its pros and cons. Even scoring systems used in pro sports are not perfect. The one I've outlined I think is at least good enough and serviceable for what it's meant to do.

[dunpeal2064]

I don't know how competitive some of the folks here are

There are quite a few members of this forum that are very competitive.

[captpain]

If you voted DoDonPachi, you're a crazy person. I felt like I should say that.

[heisenbergman]

I don't know how competitive some of the folks here are

There are quite a few members of this forum that are very competitive.

Well, I would guess that in any group of people, there would be some I guess if they want to go through the trouble of doing the number crunching to have an upper hand, then okay There are limits to how much the system can be exploited though, even if there are those who might choose to do so.

[Zerst]

Also you have to consider that some games distort score progression by quite a bit.

In Sexy Parodius, for example, a typical 1cc is about 1.2 million points. Then the game adds 1 million for each life remaining. That means someone who clears the game on their last life gets ~60% if the top player scored identically but died one time less. P1 is better at the game, but not twice as good as P2. Then next week P2 counterstops mars matrix and P1 only gets 650 billion. P1 is still ahead of P2 in the overall rankings.

[heisenbergman]

If you voted DoDonPachi, you're a crazy person. I felt like I should say that.

I feel like Gunbird 2 + Sengoku Blade would have been perfect.

But I saw prior to creating this thread that there has apparently been no DoDonPachi competition yet (or STGT at least) and figured I'd throw it in the poll.

[trap15]

(1) hold off from posting a score until the last moment so they could compute how much points they would get

This happens quite frequently in STGTs.

(2) compute their points as a percentage of the highest scorer

Terrible for the reason Zerst mentioned.

(3) determine their ranking among all participants to be able to get their final points for the week [which would mean going through the whole thread to collate the scores of the other players]
(4) compute how much of an effect those points would have on their weighted average [which would mean they have to keep track of the median score across all players to use for the formula]

Theoretically this would already be done for them, if you have a high-score list that you update fairly frequently.

I don't know how competitive some of the folks here are to be willing to go through that just to get an upper hand

I can already tell you, a lot.

Don't rely on players not to abuse the system. They will. Just look at ESP Ra.De. to see that in practice. I want to find the developer interview about that, they said they knew about the scoring trick, but they ignored it since they didn't expect anyone to actually use it.

[BIL]

They will. Just look at ESP Ra.De. to see that in practice. I want to find the developer interview about that, they said they knew about the scoring trick, but they ignored it since they didn't expect anyone to actually use it.

Page 8 of gamengai's fantastic-as-always The World of ESPRade translation.

I'll always remember that interview. Years ago it was my first exposure to the culture of this genre. Up to then I'd usually read things about Gradius V being a credit feeder, couldn't fathom someone actually keeping a boss around longer than necessary.