BuzzerBeater Forums

I disagree that 1v1 trains the same amount of HA and DR. 1v1 guards trains more HA for bigs. 1v1 forwards trains more DR. 1v1

This is wrong, you can have a look at the results from the US offsite project for confirmation.

There is no evidence that 1v1 guards and forwards are any different in respect of primary training (so HA and DR). The only difference is that the secondary training is concentrated in JS for guards and is split for forwards.

Yes guards trains more JS, but my point was why would you want to give up IS for more JS? It's almost never a good trade-off.

Cross training and elastics are not the same thing.

Again, I'm not sure what you're talking about. I haven't mentioned crosstraining. What I provided is primary and secondary training. You always get training in the skills I mentioned, crosstraining is random.

Higher IS has a positive effect, while higher SB has a small negative effect.

This is incorrect, higher SB (than ID) will make ID train faster, higher ID (than SB) will make SB train faster. That is exactly how the elastic effect works.

IS is affected by JS, OD, ID, and RB. Higher JS, ID, and RB has a positive effect, while higher OD has a small negative effect.
RB is affected by ID and JR. Higher ID is positive and higher JR is negative.

Wrong wrong wrong. OD and RB have no elastic influence on IS. JR has no elastic effect on anything except for JS.

Also there is no 'negative effect' it's a simple relationship: the higher skill pulls the lower skill making training faster. Obviously if you flip it around, you will have the opposite effect (lower skill drags the higher skill).

this is theoretical based on data put together by BB users.

Well, it was a statistical analysis based on thousands and thousands of data points. Regressions have been run on that data set in order to estimate the training speed and the impact of other factors like height, age and trainer level. So while the model built on that analysis is not exact, it is a very good approximation. Besides in respect of identifying which skill have an elastic effect on other skills the error should be basically zero. Keep in mind that the elastic effect was something the developers talked about, so people looked for it in the data: while we can discuss how accurate the elastic effect estimate from the analysis is, we can pretty much be sure of the relationship between different skills (i.e. you can be 99.99% sure skill A has an elastic effect on skill B, however you can only say that the effect is x% with a lower level of confidence).

From my experience one on one for guards is a better option to train big guys from one on one forwards.

Apart from the rest which is also questionable, but I will address later, this is seriously bugging me. Why would you want more JS on a big man? Check what it does to salary and cap: you would be paying good money in salary for the privilege of having less cap room for more useful skills!

Also JS will have a good elastic effect later on IS skill.So definitely its a better choice for taller players.

That happens because JS has an elastic effect on IS, but not vice-versa. So if you look at the total amount of skills it will always be true that 1v1 guards will give more TSP, irrespective of height, but the difference is lower for tall players as the height effect on IS and JS secondary training offsets part of the one way elastics. Of course, you will not get more IS training through elastics than you would from direct secondary training.

From coachParrot:
6'2'' player 1v1 guards then IS...https://imgur.com/yE3ibiV
6'2'' player 1v1 forwards then IS...https://imgur.com/OqJPODt

6'10'' player 1v1 guards then IS...https://imgur.com/j5aygKT
6'10'' player 1v1 forwards then IS...https://imgur.com/ZHqHIrf

Since IS is more valuable than JS at any position and since JS costs money and eats in the cap, it's actually better to use 1v1 forwards unless the player is very short, you are trying to build a high JS/JR player or you have a HoF potential and you are looking to hit the highest possible TSP (irrespective of the value of skills).
To explain what I mean regarding the impact on cap and salary, consider the following example:
- This player with 12 JS has 167k estimated salary and can likely still be trained at full speed in primaries: https://i.imgur.com/LkiY5rS.png.
- The same player with 15 JS is pretty much capped and has 209k estimated salary: https://i.imgur.com/7DImZo2.png. At PF or C, would you take this player over https://i.imgur.com/KNlwh06.png? This is the trade-off.

Last edited by Lemonshine at 11/27/2019 8:39:11 AM

1on1 guards doesnt train IS compared to 1on1 forwards which trains both IS and JS. That might be useful in a scenario where your trainee has very low JS. I do rely on elastics in training, so I can also see where you are coming from. I think every trainee is unique. When someone prescribes a generic training plan disregarding how skills are broken down it shows that elastics are not given equal weight. Every trainee is different.

Per USA off-site data, taller bigs should be trained in 1v1G because it trains HA quicker on the bigs and IS actually trains slower the taller the player is.

1v1F trains IS quicker for shorter guys.

So as a result, it is better to train 1v1F for guards ( typically 6'4" and shorter) and 1v1G for bigs (6'11" and taller). For guys between 6'5" and 6'10" I think 1v1F would probably be fine.

1v1G because it trains HA quicker on the bigs and IS actually trains slower the taller the player is.

It's like you guys have never looked at the training simulator or coachParrot which use the coefficients estimated based on that data or just go by hearsay.
1. 1v1G and 1v1F train HA and DR exactly the same way
2. The training on HA and DR is unaffected by height

IS actually trains slower the taller the player is.

IS always trains faster for taller guys. Always. Both as primary (IS) and secondary training (ID, RB, JSF, 1v1F).

As I said before, JS has an elastic effect on IS but not vice versa so if you are planning to have the most TSP, then yes 1v1G will give you highest TSP. However since it always results in less IS it's usually not a good choice in terms of salary, cap and on-court efficiency, except for a few selected situations which I already listed.





Last edited by Lemonshine at 11/29/2019 6:24:46 PM

Is it possible that the IS training speed slowed down because of negative elastics. Was that factored in your old data? I got to agree with Lemonshine on this topic. I would have to say that coach Parrot is 90% accurate for me. Maybe you can share your data with us to compare.

This is the table used in coachParrot for elastics calculations:
https://imgur.com/smjYhlI

This is the table with the training coefficients:
https://imgur.com/gm8zvYB

And these are the other tables for height, age, trainer (cap coeff was most definitely changed by me):
https://imgur.com/2YbTZLX

I now see that rhyminsimon's tool had a decresing relationship, but he also has 1v1F being faster than 1v1G by over 10% for short players, which sounds highly unlikely. The distortion in the coefficients for HA and DR is likely due to skills that affect HA but not DR via elastics (OD). If you look through his coefficients (which were older than the ones Jozef Ka worked with) there are a lot of things that make little sense. If you look at IS training at 6'0'' he estimates the secondary training on JS to be 3% of primary training, while at 7'6'' he estimates a 30%.

I would need to see the dataset and run my own analysis, but at a glance I'd say rhyminsimons numbers defy common sense. According to that, JR and SB benefit from no elastic effect, so yeah.

Last edited by Lemonshine at 11/29/2019 10:07:18 PM