[NoOne]

Revisions likely - spreadsheet will be added later. a version of it is in the forums. no idea what i've changed since then, but 95% should be the same. go ahead and use that until i post a new one. search my posts it should be titled League Total Modifier or the like.

Edit: a spreadsheet with 100years of data and how to use it.. many comments throughout. it's 1 step from final simulation to verify. plus, i put soem screenshots in of the spreadsheet.

***************
if you prefer to use auto-calc, this is not for you. this thread is not a philosophical debate.

All important game settings must be set as you intend to use them. even scouting accuracy and the scale you intend to use. it WILL make a difference. things like awards, news, bb cards, facegen etc do not influence the game in any way, so they are not needed for testing purposes. if you are not sure of its effect, leave it as you will use it.

the spreadsheet will work for 2 sub-league Leagues with the same rules. otherwise, you will need to make your own or re-work mine.

Ignore the Hits LTM suggestions from the attached spreadsheet. it is wrong. -- its calculation is explained in the process below.

Assuming you have set your League Totals or are happy with defaults. it doesn't matter what you use, as long as you type it into the spreadsheet and it matches the League Totals in the game. if you don't want the resulting slash line from default League Totals, i suggest you come up with your own league totals. calculate BA, obp, slugging from the totals... make them as you want the results average.

Gidp and Lidp may or may not be right in my spreadsheet. ootp wants very few liner dps - assuming feedback in csv data dumps is reliable data. it shows total dp and gdp.. so i assume the difference is LiDP - but some could be K-throw'em outs too. i assumed that the liner dp was far more common than a strike'em out - throw them out DP.

BABIP / BB / SO must be done first... step 2,3,4 may be better in a different order. not sure. this is what i did.

***************

Step 1: set some base LTM. Create initial Long-Term simulation data.

auto calc 10x <threw up i my mouth a bit> and take the averages. unless your league is perfectly average talent-wise in year one, this will not be enough. however, it gets you in the ballpark for sure.

if you did not adjust LT to match the slash line you want, autocalc will not work well in this process. if you are happy with default or you have input adjusted totals, it will work well.

Start the simulation but take it slow the first couple years. merely look to see if something is MASSIVELY wrong... do not micro manage at this point. knowing ranges of various stats will help. i will post some guesstimates for you below based on a 28T/156G league - you may need to translate those values.

e.g. let's say hits can range 3000, so +/- 1500 from *expected* average is normal. so, don't get your panties in a bunch if hits are 1500 too low/high the first year or two. if it is in the expected range, do not tweak!! you cannot know the average from small samples, but you can easily see if they fall in an acceptable range of values. if you start tweaking now, it will be endless, and you will never reach your goal.

If all looks good, let it run wild for a long time. do not micro manage!!


Step 2: Calculate new Hits LTM from the long-term sim created in step one. Hits LTM affects BABIP!

Example Caclulation: BABIP is off by .002 and the league averages 150k AB and 30k hits.

.002 * 150k AB = 300hits. so, we must lower/increase current average hits by 300.
300/ 30k hits = .010 so, we must lower our Hits LTM by 1%. (please note that .900ltm * 1% does not equal 1.000ltm * 1% -- so, .010 does not mean deduct .010!)

Enter the new Hits LTM you have calculated and ...

--you can also implement suggested LTM from spreadsheet for: HBP, GB%, WP, Balks, PB, SF, SH, SB, SB%, GiDP and LiDP. (i could be wrong about sb and sb% - 50/50 on those 2) from now on, these should all remain static throughout the process. No minor tweaking until the end.

-- hold off on 2b, 3b, hr, bb, and SO this round.. DO NOT change these, yet. the original auto-calc values will be fine for now.



Step 3: Verify the new BABIP and run a new long-term sim.

run ~10 years, maybe just a couple, and make sure the new BABIP is what you expect it to be - not the average necessarily with so few years of data, but all data points should fit in the range* of the new BABIP baseline. adjust as necessary (range = .014 or +/- .007 from average)

Now, continue the sim until you have a sufficent Sample size to calculate a good LTM for BB. if you didn't have to make any changes, you can include those initial verification years, otherwsie exclue them if any changes were made. this data will be used in step 4.



Step 4: Setting BB LTM and verify.

after 50 years the average will still be a bit wild, fyi. you are better off with more years. once you set this value, do not move on yet... sim 1-2 years and verify the results are in an acceptable range - just like before with BABIP. average is not important for 1-2years, so only focus on whether it is in the expceted RANGE. all is good?


Step 5: Set SO LTM and verify.

try this with the sim data used in step 4... run a year and see if consistent. if not, you will need a new long-term sim with the new BB LTM from step 4. some may need it, some may not. these 2 LTM affect each other greatly.

Do the same with SO as you did with BB. "SO" likely require less years than BB to level off. maybe 50 is good enough? once you have a long-term average, set it and verify as you did with BB. all good? within RANGE? move on.


Step 6: Work on 2b, 3b, hr

as with step 5, if you use data from step 4 and it works, do it. otherwise you need a new long-term sim with the new bb/so LTM. you most likely will need a new long-term sim.

since SO and BB are now set, you will need to hone these last LTM a bit before the next Long-term simulation. they should not be far off. if they are all very close you could just run a long-term sim immediately and skip the rest of this step. otherwise, we will slowly adjust one at a time based on the spreadsheet suggestions. don't start with 3b.

change and verify as we did with 1-2years of data for SO, BB and babip. as long as it reacts as expected, move to the next LTM to chnage. if you change one and it raises/lower another, you will have to apply a little art to the science. guesstimate how much it affected that average... make adjustment from that guesstimate.

Once they are all relatively set, you need to run a long-term simulation for final tweaks. as long as you dont adjust hits, SO or BB things should move along smoothly in this step.

Run a new long-term simulation. exclude any initial years in which you made changes. you need it to be long enough for the most volatile LTM statistic to level off... how many years? no idea. i'd start with 75-100 and be willing to ignore minor differences.


Step 7:Analyze sim data. make minor tweaks, if necessary.

at this point, Do ONE TWEAK AT A TIME!! verify that results remain in expected range before moving to next LTM tweak. if nothing changes wildly beyond expectation, it is very likely going to do what you or the spreadhseet calculated.

if you change anything drastically or just unsure, run a 50-year test... realize that if something is off be less than 5% it's probably fine with this smallish sample size. min/max values and range will help you in verification, too.

__________________________________________________ _______

General tips/tricks/nuances:

apply common sense... e.g. let's say after all doing all of the above everything is great except 3b is off by 10. changing 3b isn't helping and causing other LTM to get all messed up.... Keep It Simple (stupid). 10 3b over 30 teams and 162 games is nothing. it won't even register a .001 change to Slugging Pct. if it is close, be happy!@$#@ it's something i struggle with, lol.

do one change at a time when told to do so: ltms and results are not 1:1 at all times for all LTMs. it may suggest to lower it .020, but due to other changes you may actually need raise it .020 points. i am fairly sure the ones i tell you to set asap are mostly independent from other LTM except in extreme cases (again, only 50/50 about SB attempts and sb% -- these can be a pain and they swing wildly year-to-year so you need a HUGE sample size to do them precisely).

so, if you change one LTM and the results of another swings wildly when you verify the results of the change, you'll likely need to use a different starting point to calculate that other LTM that change dratsically - or a whole new long-term sim. apply common sense. try to avoid it if possible. either way you'll be working with a new average when this happens and you must adjust. unless it's near the end of the process, ballpark it and see if it works short-term. or you can do a long-sim to learn it with more certainty. if it's in the ballpark, the final long-term sim near the end of the process will likely fix it for you.

Save old spreadsheets with long-term data... referenceing them may be helpful when odd things happen. LTM1, LTM2 etc.. named spreadsheet files.

do all testing on the same league you intend to use. you can turn news and all the frills back on once you're done testing. do not turn off anything tangible that you will use. e.g. if you use owner goals, keep it on. missing one of those settings will not wreck your stats output. Whereas missing an important setting that causes your results not to match your test runs is a big problem. maintain a backup at the year you intend to start the league. all testing can just keep running... erase history/stats as needed etc.

If you erase league history/stats AND you use AI Eval, some things may not match up correctly the first 3 years. E.G. SB will slowly ramp up. these AI decisions obviously require history/stats to work correctly. so, the first 3 years SB attempts will be low. mine went 1000successful SB to 1700 to 2500's year 3.. the percentage looked fine from year 1. i don't think i noticed anything else odd... when i erased history/stats i simply remove/ignore the first three years of any test-simulation.

During one of the Longer sims where the LTMs and results are close to what you want, start noting the Range of results for each LTM statistic. knowing hits can range 3000 and SB can range 1000 helps. you know if hits are within 1500 of the average you want, the changed results are falling within an acceptable range. this way you can do a simple verification of a change without knowing the resulting average... it gaurantees nothing... it just helps avoid obvious problems.

Also, knowing the range can help you guess the average before you have a solid average. you can see if the data is skewed to one side of average. if the diff between minimum and avg is larger than max and avg, you can guess that there are more data points on the higher side. don't do this wiht a few years.. if you don't have an accurate high/low mark it won't work well. you could make some charts with distribution curves - that would help immensely.

************************************************** *******************

For Seach: league leagues league's total totals modifier modifiers how what howto how-to auto-calc autocalc