Testing WhatIfSports League (TWISL)

Section 1: LEAGUE PHILOSOPHY

The TWISL is a multi-season league in which each season is designed to help us better understand the SLB algorithm. Each season will be different, and owners may come and go as they choose depending on their interest level in each season’s theme. There is no expectation that owners commit to multiple seasons.

However, it is important to know that this league is designed first and foremost to create better understanding of how SLB really works. This means that each season will involve various roster constraints, or theme rules, and that occasionally these may result in one or more teams being at a disadvantage compared to others. While I will do my best to construct rules that will allow each team to be as competitive as possible within the construct of each season, that will be a secondary consideration. If you signup to participate in any season of this league, please do so with the understanding that the playing field may not be level.

Section 2: LEAGUE STRUCTURE AND WIS RULES


TWISL is intended to be a 24-team league each season. If we do not have sufficient owners to reach 24 teams within a reasonable amount of time after the first theme posting (typically 7-10 days) then we will use a smaller league (16 or 20 teams, depending on how many owners signup). This will not be put to a vote. Please only signup if you are willing to play in any format (16, 20, 24 teams).

While the exact WIS rules for each season may vary, in general each season will have an 80M cap, no DH, no AAA, no WW, no trading, with injuries ON. Clones may or may not be allowed depending on the specific theme idea we are playing each season.

Section 3: THEME IDEAS

TWISL will last as long as there are interesting questions about SLB for us to test in a workable theme that will attract owners. Some ideas for specific themes will be:

• What is the effect of pitcher K rate on range (plus plays)? THIS WAS THE THEME FOR SEASONS 1 AND 2
• How much does fatigue impact hitter performance?
• How much does fatigue impact pitcher performance?
• What do park factors really mean? How much additional offense is generated by a park that is +1 or +2 for 1B/2B/3B/HR compared to a neutral or negative park?
• How do the park factors work in combination? Will more triples be hit in a park that is +1 or hits and +1 for triples, or in a park that is -3 for hits and +3 for triples? THIS IS THE THEME FOR SEASON 3.
• How important are manager settings? If you set all of your base stealers to 4 or 5, how many more SB attempts will they have compared to setting them to 1, 2, or 3?
• What impact do the H&R and baserunning aggressive settings have?
• How do the LH and RH specialist relief settings work? What is the impact of setting all your short relievers to LHS/RHS rather than (say) Setup A/B?
• What is the impact of range on double plays?
• There are no doubt many more, and owners are encouraged to suggest their own ideas for us to test. The only condition is that it has to be something around which we can design a workable set of league rules.

Note that to make the tests as reliable as possible, for most of the above we will need to put significant restrictions on roster composition. For example, if we are testing the effect of baserunning aggressiveness, we will likely want each team to have exactly the same set of offensive players (or at least the same starting lineup). Otherwise there are significant confounding variables that will dilute our ability to understand the specific effect we are looking for.

OK folks...I am just about to start signups for season 3, but first I would like to crowdsource some input from the WIS braintrust.

I am going to require each team to use the same 8 position players:

		Pos	Player	B	PA/ 162	AB/ 162	HR	RBI	SB- CS	SO- BB	AB/ HR	AVG	OBP	SLG	F/R	Salary
		C	1885 Charlie Bennett	R	594	524	5	60	24-21	37-47	69.8	.269	.356	.456	C-/A+	$5,000,150
		1B	1894 Roger Connor	S	647	567	8	93	19-26	17-59	57.8	.316	.400	.552	D/A-	$4,693,949
		2B	1954 Pete Runnels	L	600	514	3	56	2-3	60-78	162.7	.268	.368	.383	C/D	$3,904,374
		3B	1944 Bob Elliott	R	654	570	10	108	9-10	42-75	53.8	.297	.383	.465	C/B-	$5,069,467
		SS	2006 Jose Reyes	S	703	647	19	81	64-17	81-53	34.1	.300	.354	.487	B/D+	$5,625,530
		OF	1905 Elmer Flick	L	603	526	4	64	35-30	46-53	125.0	.308	.383	.462	D/D+	$4,598,146
		OF	1912 Chief Wilson	L	690	626	11	95	16-14	67-35	53.0	.300	.342	.513	C/A-	$5,708,292
		OF	1985 Willie Wilson	S	642	605	4	43	43-11	94-29	151.3	.278	.316	.408	A/A	$5,740,846

Each of these guys has a 3B/100# of 3 or higher...including the MLB single season record holder for triples, 1912 Chief Wilson.

But there are two things I am grappling with:
-- How many different parks should we test? At minimum I think we want to test 2 parks...one that is positive for hits and either neutral or negative for triples, and one that is neutral or negative for hits and positive for triples. But do we want to include more parks than that? Should we have WIS Park (zeroes across the board) as a control? I could see an argument for allowing 2, 3, 4, or 6 parks; whatever we decide on, there will be an equal number of teams assigned to each park.
-- Should I pre-select some or all of the pitching staff for each team? Or let everyone choose their own? Obviously the benefit of pre-selecting is that our test will be purer; the benefit of not pre-selecting is that owners get more control over their rosters and it may make the league more fun and competitive for everyone. I could split the difference perhaps and maybe select 800-900 IP that everyone must use, and then let owners fill out the rest as they so choose?

Please share your thoughts...and more importantly, share WHY you think that way.

I think 2 parks would be best. We want to increase the sample size and that would be accomplished best by limiting the number of parks. This way we would have 12 teams in one park and 12 in the other. If we do 3 parks then there will be only 8 teams per park. Let’s go with the larger sample size. Just my opinion. As for rosters I like standardizing as much as possible. I like as little individual input as possible so the results don’t skewed too much by a few well selected anomalies. I will join either way. I enjoy your leagues.

I don't think we need a neutral park here. A simple A/B of +/- and +/+ will be the most effective way to measure the difference.

How are you thinking about alignment? One type of stadium per conference?

For pitchers, we should consider guidelines. I don't know how prescriptive we need to be.

I agree with two parks. Maybe do one league with one park, one with the other, so you really see the differences?

I think setting a required pitching salary range and inning range might suffice, but let everyone pick their own staff. We should get enough variety to yield a balance for results purposes while allowing the owners something they get to "own" in the decision process.

Another thing that we may want to test at some point is the effect of pitching ERA (meaning, range of years, not earned run average) has on XBH.

So I would say we probably need to specify a range of years (say 1920-40, or 1990-2019, or whatever) and let people pick their pitchers in that range.

Also, booooooo on no Curtis Granderson!

Alllllso, if we're testing an offensive theory why don't we use the DH? More data points!

agreed on limiting pitchers to a specific era, based on anecdotal evidence where modern triples specialists seem to get more triples in open leagues (where most of the pitchers are deadballers) than triples specialists from the deadball era, seems like all pitchers should be from, say, 1910-1919 or 2010-2019 to control for any potential difference that makes. if you do an 80 million cap that will leave some flexibility for everyone to not have to roster the same stud pitchers.

also agreed on two ballparks only with each park unique to each side of the league.

Posted by redcped on 3/9/2020 11:42:00 PM (view original):
I agree with two parks. Maybe do one league with one park, one with the other, so you really see the differences?

I think setting a required pitching salary range and inning range might suffice, but let everyone pick their own staff. We should get enough variety to yield a balance for results purposes while allowing the owners something they get to "own" in the decision process.

I'm thinking the best way to handle pitching might be a little weird.

If each conference is made up entirely of one park, we should have each pitching staff represented the same amount of times in each conference. So, at minimum, everyone has the same staff, but at maximum (and this would be the weird but fun option) each team has exactly one exact copied pitching staff in the other conference. 24 owners = 12 staffs

Thanks for all the good input guys - been a crazy time at work, as I expect it as for many others. I will try to get this set up this weekend.

Hope everyone is staying safe.

Signups are now open. See separate thread. Thanks for the input guys. jfranco, I have bowed to the peer pressure and swapped Granderson in for Elmer Flick....

Please do NOT sign up in this thread - see separate thread to sign up